In 1978, René Coelho founded MonteVideo in Amsterdam as the first gallery in the Netherlands dedicated to video art. The initiative emerged as portable cameras, electronic image processors, and television monitors were giving artists new ways to produce and distribute moving images outside conventional film and broadcast systems.

Pioneers of Media Art: The First Years of MonteVideo, an exhibition at Nxt Museum, examines this formative period through artworks and archival material preserved by LI-MA. Presented across ten spaces, the exhibition connects early experiments in video, animation, light, and electronic sound with contemporary audiovisual installations.

Developed in collaboration with LI-MA, the exhibition opened to the public on June 12 and remains on view through September 10, 2026. Central to the presentation is the work of Livinus van de Bundt and his son Jeep van de Bundt. Livinus’s experiments with electronic images directly influenced Coelho’s decision to establish MonteVideo. The gallery opened with an exhibition centered on his work, including Moiré, his collaboration with Jeep.

Building an Infrastructure for Video Art
MonteVideo operated as more than an exhibition venue. It provided artists with access to production equipment, collected and distributed artworks, and helped establish video as an independent artistic form. These functions were particularly important because video art depended on technologies that were expensive, technically specialized, and rarely supported by conventional galleries.

The institution’s early program included artists working with television, electronic image processing, performance, sound, and image manipulation. The Nxt Museum exhibition presents works by Nam June Paik, Nan Hoover, SERVAAS, Tom Klinkowstein, and Trevor Batten alongside those of Livinus and Jeep van de Bundt. Together, their works examine television, electronic image processing, performance, transmission, and the relationship between sound and moving images.

MonteVideo’s institutional development also helps explain why much of this material remains accessible. In 1993, MonteVideo merged with Time Based Arts. The organization later became the Netherlands Media Art Institute, or NIMk. After NIMk ceased operating at the end of 2012 following funding cuts, LI-MA was established in 2013 to continue work in media-art distribution, preservation, research, and presentation. LI-MA’s collection now incorporates material from MonteVideo, Time Based Arts, NIMk, and other Dutch media-art organizations.

Generating Images with Sound
Livinus and Jeep van de Bundt’s Moiré (1975) provides a technical and conceptual anchor for the exhibition. Livinus, originally trained as a painter and graphic artist, began experimenting with video around 1970. Jeep, a musician and percussionist, contributed electronic compositions and explored methods for connecting sound directly to visual output.

In Moiré, electronic audio signals generate and distort abstract forms on screen. The artists used a Pong console, a tone generator, manipulated television circuitry, photographic slides, and a self-built image-processing system called the Lumodinamiese Masjiene, or lumodynamic machine. They altered the convergence of a television’s red, green, and blue channels to produce displaced colors. Other sections were created by photographing images from a monitor and filming the resulting slides frame by frame on 16mm film.

The six-minute work connects sound and image through a shared electronic process. Rhythm does not merely accompany the visuals; it contributes to their movement and structure. Although completed in 1975, Moiré premiered on February 25, 1978, as part of MonteVideo’s opening exhibition.

For the Nxt presentation, LI-MA worked with Archival Consciousness to digitize a selection of Livinus van de Bundt’s experimental video works. The process addresses a recurring problem in media-art conservation: preserving an artwork may require transferring it between formats while retaining the visual, sonic, and temporal properties produced by its original equipment.

From Analog Circuits to Spatial Media
The historical works appear alongside contemporary installations by 4DSOUND, Balfua, Boris Acket, Children of the Light, Lumus Instruments, and YONK. These projects employ spatial sound, synchronized lighting, kinetic systems, and digital imaging. Their inclusion extends questions already present in MonteVideo’s early program: how technologies structure perception, how signals become artistic material, and what happens when an artwork’s technical platform becomes obsolete.

Pioneers of Media Art presents MonteVideo’s history as part of a longer technological and institutional development. Its subject is not simply the adoption of video as a new artistic tool. It examines the network of artists, equipment, galleries, distributors, archives, and preservation specialists required to support an emerging medium and make its history accessible decades later.

You expect a raid. A bot built by the artist duo !Mediengruppe Bitnik buys ecstasy from the dark-web marketplace Agora, and a museum hangs the parcels on the wall. Instead the show runs its full three months. The bot buys in plain sight; the post carries the contraband to the Kunst Halle Sankt Gallen as ordinary mail. The artists warn no one; they let each rule-system run to the letter. The show closes on schedule, and in January 2015 the St. Gallen prosecutor seizes the work. Months later he clears the artists and the bot of every charge, ruling in writing that the public interest justified exhibiting the drugs themselves as art. He returns almost everything (the counterfeit jeans, a scanned Hungarian passport) and destroys only the ten ecstasy pills, not for failing the art-test the ruling had just granted but because the police would not release an object that could be swallowed. Five years later the state's own collection acquires the work. The art-claim is conceded in full, in writing, and the hand falls on the matter regardless. Call it compliance as material: not the parody that mocks the form from outside, nor the refusal that walks away from it, but taking the form itself as the medium, executed with the fidelity it demands, until the form, and not a picture of it, is what stands in the room.

The obvious frame is overidentification, embracing a system's forms so straight-faced that the obscenity it hides rises to the surface. But overidentification speaks in the artist's voice: the excess is the artist's, Laibach over-performing the state until the performance accuses it. Here the accusation is in the regime's own voice. The artist authors only the conditions; under exact compliance the regime authors the position, and that compelled self-description is the work.

Benjamin Buchloh's "Conceptual Art 1962-1969: From the Aesthetic of Administration to the Critique of Institutions" (October 55, 1990) named a related inflection: artists adopting bureaucratic registers. But Buchloh's registers were authored by the artist; this work operates inside externally-authored scripts: state law, platform terms, evidentiary protocol, the public-art contract. Maria Eichhorn's Aktiengesellschaft (Documenta 11, 2002), a self-owning corporation incorporated under German law, is the cleaner hinge. After Eichhorn, the door is open.

The works are not mine to claim; the pattern between them is what I am naming. Aksioma in Ljubljana has shown many of them across a decade, through Bitnik and Moll, through Tactics & Practice and Are You a Software Update?, without naming what runs between them. They divide on one question: does the piece force the regime to write its own self-description? Where it does, that document is what the show leaves behind. Goldin+Senneby's Eternal Employment (Korsvägen, 2017) is a state-issued contract paying one person to clock in at a Swedish train station for its projected lifetime; in April 2022 the Swedish Transport Administration cancelled the commission it could not finance, and the cancellation, not the contract, is what it produced. Cameron Rowland's 91020000 (Artists Space, 2016) makes the institution open an account with Corcraft, the state's prison-labour supplier, and buy goods made by incarcerated workers; the work's title is that customer number, and what survives is the paperwork. Each, like Bitnik, forces the regime to leave a record in its own hand.

A second mode runs lawfully inside a system to expose it, without compelling any reply. Joana Moll and Tactical Tech's The Dating Brokers (2017) buys a million dating profiles from the broker USDate for 136 euros, a year before GDPR, the last legal window of a market the regulation would close. Tega Brain and Sam Lavigne's Offset runs the voluntary carbon-market grammar to the letter, issuing tradable credits for industrial sabotage booked as temporary carbon storage, the Tyre Extinguishers' deflation of thousands of SUV tyres among them. Constant Dullaart's fake Facebook accounts, named for eighteenth-century Hessian mercenaries, satisfy platform authentication and nothing else. The compliance is total, but here no one is compelled to answer, and though each artist has named some part of what they do, none has set the whole tendency under one heading.

In the first mode, the regime's own document outlives the show, real and so transmissible, the event's own paperwork rather than a depiction. Critics will call this operation, not critique, and so toothless. The sharper objection is the reverse: that faithful compliance flatters the regime, the prosecutor who spares the artists looking liberal for free. But the magnanimity is not free. The letter has to be written, and writing outlasts the occasion, fixing on the record what the regime would rather keep tacit: that it granted the art-claim and fell on the object anyway. And these works depend on institutions willing to host them; that dependence is the subject, not the alibi, since the compelled document marks where the hospitality ends. The grammar that demands to be filled also decides who can fill it and who is shut out before the question of art arises, and the work that masters it exposes the threshold at the door.

We click Accept many times a day, filling forms we did not write and cannot edit. These artists fill theirs on purpose, to the letter, until the form gives up what it is for. The exhibition already exists, scattered across a decade and held in places like HEK and iMAL; what it has not had is a name to argue over. Compliance as material is one. The bet is not that these artists are stronger than the regimes they enter. It is that they have read the rules more closely, and the sharpest of them have made the regime put it in writing.

Indian computational media, cultures and even subcultures have becomeIndian computational media—and the cultures and subcultures that have emerged around it—have become an important site for examining a methodological paradox. Home to one of the world's largest populations of engineers, technologists, and AI researchers, India has developed substantial digital capacity while simultaneously revealing significant questions about the infrastructures that support computational art, knowledge production, exhibition practices, and cultural exchange. Here, "synthetic" refers not to artificiality but to infrastructures assembled through improvisation, DIY cultures, lived experience, and practices that resist disposability.

These technologized refractions are profoundly global in their circulation of ideas and tools yet profoundly rooted in contextualization and aesthetics. There exists an intricately woven ecology of international embassies providing patronage, outreach partners, global corporates, design, engineering, art, and film schools with their new media festivals and degree showcases, newly formatted techno-media conferences, virtual worlds, media labs, SMEs, informal collectives, artist studios, hardware platforms, exhibition architectures, and labor formations which range from housewives with forehead rigged cameras recording peeling of mangoes to Tate Modern’s recent artistic showcases.

Academia and industry have always tried to help–and exploit–each other. Yet some of the most conceptual computational works almost always arrive from this friction between the two. Rayscapes Arcade (2025–ongoing) is one such proposition. It was conceived by the transdisciplinary collective Coded Playhouses, comprising academics, filmmakers, architects, designers, medical doctors, and creative technologists. The project emerged from a provocation by film scholar Dr. Kaushik Bhaumik, who has taught a course on Satyajit Ray at the School of Arts and Aesthetics, Jawaharlal Nehru University, for more than five years. The question was deceptively simple: what would it mean not merely to conserve Rayscapes Arcade's world, but to computationally expand it? The outcome unfolds through 4 quadrants:

1. A playable archive of Sound and Memory (Unreal + Touch Designer Game)
2. A Cartography of Rayscapes Arcade's Cinematic Geographies (Architectural City Model)
3. A Zoetrope reauthoring Rayscapes Arcade's Archive into Living Memories (Optical Toys and Textiles)
4. The recreation of Rayscapes Arcade's compositions via contemporary imaging technologies (iPhone and Dji Drones)

If Rayscapes Arcade computationally expands the archive,यथा वणिजो लाभकामाः समुद्रम्-Like Merchants Desirous of Profit Who Venture Upon the Ocean (2026) directly takes on the infrastructure head-on. Built partly in Touch designer and Unreal Engine and currently on public viewing at British Council, New Delhi traces an arc between colonial telegraph wires, Undersea Cables, Wave Powered AI Inference Platforms in the Global South. The work showcases the Anthropocene’s most honest image and anticipates what’s to come for India’s land, water, and air, alongside the further selling of its elements to new merchants (Meta, Microsoft, Open AI, X, and related actors).

Yet all these formats and modes of curation have a great common factorial – a structural signature of an archive remix, Indian living traditions technologically recoded and heritage conservation enveloping the canvas.

ΨIndia = ∑ᵢ₌₁ⁿ (Ti ⊗ Ai ⊗ Li ⊗ Si) + ε – η

Like Unreal Engine’s Niagara particle simulation, India’s computational cultures can be represented through an equation in which ‘T’ denotes Living Traditions, ‘A’ denotes Algorithms, ‘L’ denotes Labor, ‘S’ denotes Synthetic Infrastructures, ‘E’ denotes Emergent forms, and ‘n’ denotes Ethical Friction. Yet ethics remain external to the equation, appearing as the friction that determines whether emergence becomes emancipation or extraction. 

But 1+1≠ 2 as every encounter changes the identity of the terms themselves.

Furthermore, and this is an interesting framework to follow here, as Refik Andol’s DATALAND is now open to the public and global art and film festivals simultaneously champion alternate reality pavilions, the more pressing question becomes about figuring out the right metrics to judge a computational artwork. What happens to these artworks when the grant producing the exhibit is finally executed? Does the blockchain solve the provenance issue? Who owns the archive here, and majorly who consumes these works? Who commissions them? The idea of access and accessibility becomes a beat wherein the image changes, but the infrastructure does not. 

It's important to note that various new media software and hardware labelled as open source, marketed with student discounts and tag lined as affordable for everyone are precisely built upon the miniaturization of labor. So, what’s the alternative here if not the tokenization of the entire system and outsourcing hardware which is powerful enough to run them all at 24FPS, or whichever standardized format listed by major art shows. This essay is, in essence, a call for an ethics and aesthetics manifesto wherein the task before us is not to merely build new worlds but to rather engineer conditions under which worlds can remain plural, reparative, and sovereign – for if 1+1≠2, then every computational permutation and combination produces a new reality, and the politics of that reality must be ours to shape. 

The tensor product (⊗) indicates the co-constitutive and entangled relationship among these domains rather than a multiplicative operation in the strict mathematical sense. The summation (∑) suggests that these interactions occur across multiple instances or sites (i = 1,…,n).

Additionally, the question of archive is perhaps most urgently raised by practices such as Elsewhere In India whose work utilizes archival restoration, digital reparation, gamic universes and DJ sub-cultures of remix and sampling. As Avinash and Murthy rightly point out that DJ and VJ worlds were built on bootleg fragments, and informal acts of cultural recombination. But what becomes of remix once every image is born digital and every archive is susceptible to algorithmic manipulation and generative intervention? Who owns the archive? Who owns the agency to alter an archive? And what forms of copyright, provenance and other IP laws become essential in a country such as India, where legal and ethical frameworks for computational media remain significantly under- planned and underdeveloped. The archive is no longer simply a repository of the past; it’s a newly curated site, a constant contestation for future world building.

The blank page is never actually blank. Designers do not solve problems so much as frame them; the framing always precedes the solving. Before a single line is drawn, we have already decided what counts as the problem, whose experience is worth designing for, and what a good outcome looks like. Victor Papanek made the same point more bluntly: the designer brings a complete cultural program to every brief, whether they acknowledge it or not. So where does that leave open-ended imagination in the age of scientific discovery? The last time humanity expanded into new worlds on this planet, it took courage and madness, but the hypothesis was always extraction framed as progress. This pattern became infrastructure, and infrastructure, once built, is very difficult to undo.

The contemporary space sector frames itself as technical, neutral, future-facing. But the systems being exported are familiar: military logic, geopolitical competition, extractive ownership models. Space is no longer a blank slate for our imagination. Even science fiction is inventing inside a box of assumptions. What gets built in space will not stay in space – the infrastructure decisions being made now will calcify into the architecture of whatever comes after..

Discovery vs Design – a Slippery Slope
Every act of discovery involves a framing decision about what is being found and for whom. Geographers and historians of science call this the “view from nowhere” problem: the claim to neutral observation that is always, in practice, a view from somewhere specific. When European cartographers mapped unfamiliar territories, they did not describe what was there – they described what was legible to their instruments and useful to their patrons. The coastlines were precise. The interiors were filled with projection and myth. Bruno Latour argued that scientific facts are not discovered but constructed, assembled through networks of instruments, institutions, and interests that determine what counts as evidence. Design operates the same way. The moment a design team defines a user, a problem, or a constraint, doors close.

Gehl’s observation about 20th-century urban planning is precise: cities were designed to look coherent from above, not to be experienced by a body at street level. The result was environments legible as diagrams and hostile as places. The car became the norm not because it served cities best, but because it matched the scale at which planners were working. Tony Fry calls this “defuturing”: design that forecloses possibilities it cannot see. A lunar habitat designed from Earth carries the same structural problem – the gaps in experience get filled with assumptions, and the assumptions are invisible to the people who hold them.

Space Imagination as Unexploited Resource
Space has always leaked into everyday life. Gore-Tex and freeze-dried food are well-known examples – materials and technologies refined under extreme conditions that turned out to be useful back on Earth. But we have failed to use space as what it actually is: the one design environment where inherited defaults have no physical foundation. Where the floor, the 24-hour clock, the territorial border, and the hierarchy of who decides are not given conditions but active choices. That is an extraordinary design resource. It is almost entirely unexploited.

Science fiction has always understood this. The genre has functioned for over a century as a laboratory for ideas too disruptive to propose directly – a safe distance from which to model alternative societies and ways of being human. Le Guin’s anarchist planets. Butler’s post-apocalyptic kinship systems. These are not escape. They are speculation made rigorous.

The practice has direct commercial precedent. Before Minority Report, Steven Spielberg convened architects, technologists, and writers to design the world of 2054. The result was not a film aesthetic but a prototype lab. Gesture-based interfaces, autonomous vehicles, predictive policing: technologies now deployed or in active development. The speculation was serious enough to become a reference point for actual product design.

Design fiction as a formal practice constructs futures as proposals, making alternative systems tangible enough to be evaluated, debated, and sometimes built. Space is the ideal setting for this work. No gravity to anchor inherited assumptions. No existing users to design around. No legal framework solid enough to foreclose alternatives. The conditions for genuine speculation are already there. The speculation, mostly, is not.

What is actually happening in space design is the opposite of speculation. Habitats are organized around efficiency and survival protocols developed for short-duration missions. Suits are engineered around a historical average of the astronaut body: male, able-bodied, within a narrow size range. Schedules run on Earth time because mission control runs on Earth time. Governance follows sovereignty frameworks that assume territorial ground. The design brief, implicit or explicit, is: reproduce what works on Earth under more extreme constraints.

Meanwhile, the questions that space’s actual conditions make possible go largely unasked. How do you organize collective time when there is no sun? How do you design for a body that has lost its gravitational reference? What does belonging mean without territory? What is a home when the environment outside is lethal? They are not in the brief.

The Human Brief
In The Universe Is an Inverted Womb, Daya Ventures and AFRY Engineering attempted a different starting point. Rather than survival, we applied soma design: treating the body’s felt experience as design material. This meant starting not from what a body needs to remain operational, but from what it needs to be human.

The questions that followed were not engineering questions. What mementos would you bring if you had to leave Earth forever? What is the smell of home? How can we avoid violence if everyone is crucial? How would we coordinate work if time didn’t exist – which, in space, it really doesn’t. How do you carry grief in a closed system with no return? What if clothes could adapt to feelings?

The assumption that those who discover a new environment automatically produce new thinking is precisely the assumption that produced every colonial project in history. The frontier is never blank. It is always a prototype in the making.

We once looked at pictures. Then, with the advent of computer vision and machine learning, pictures started looking at us. By the award-winning artist, filmmaker and thinker.

Today our world is under the watchful and tireless eye of computer vision, with cam­eras and monitors tracing our every move. Furthermore, generative AI is now able to render a synthetic world indistinguishable from reality for us to explore. Trevor Paglen goes in search of the ways and means of understanding this new visual universe. Instead of asking what these technologies “say” about the world, he teaches us to ask what they “do” and where such images come from.

Exploring the esoteric worlds of psyops, UFO imagery, magicians, and public relation gurus, Paglen shows that this appar­ently alien realm is more human, but much stranger, than we imagine.

Eric Lu, founder of AI typography platform Mixfont, has released Ghost Font, an experimental motion-based typography system that investigates a growing question in human-computer interaction: can written communication be designed for human perception while proving difficult for current AI vision models to interpret?

Despite its name, Ghost Font is not a conventional font file. Instead, it generates short videos in which letters emerge only through motion. Thousands of black-and-white dots fill the frame, with dots inside each letter moving differently from those in the surrounding background. The result is a message that human viewers can typically recognize while the animation is playing but which disappears into visual noise when paused.

The project also incorporates a decoy message intended to mislead AI systems attempting to identify hidden text. According to Lu, many current multimodal AI systems primarily analyze video by sampling individual frames rather than reasoning directly over motion, making Ghost Font difficult for many leading models to interpret without additional guidance or custom analysis. (mixfont.com)

Lu is careful not to present Ghost Font as an encryption or security tool. The project documentation notes that a sufficiently capable computer vision pipeline using optical flow or other motion-analysis techniques can recover the hidden message. Instead, Ghost Font is positioned as an experiment exploring differences between human and machine perception. Lu also suggests the project could serve as a way to observe how future video-native AI systems improve at interpreting motion.

The project also revisits ideas explored by Sang Mun's 2013 ZXX typeface, which attempted to prevent optical character recognition by disguising letterforms with visual noise. Lu argues that modern AI models can readily read static adversarial typography, prompting a shift toward motion as the primary carrier of information.

AI feels like light. A prompt entered, an answer returned: the instantaneous interface of artificial intelligence seems frictionless, immaterial, weightless. And yet the infrastructure behind the digital is often material, even geological. Behind every instant answer lies a vertical extraction axis: rare earth and critical minerals pulled from deep underground; data centers consuming water and energy at unprecedented scales; orbital satellites with remote sensing technologies reaching into unexploited territories, reorganizing matter at a planetary scale. The accumulation of data in unprecedented quantities requires further infrastructural expansion in storage, and data labeling workforces to train the system. These accumulation technologies layer onto existing ones, intensifying extractive logics rather than replacing them. The more data collected, the deeper the extraction required to sustain it.

There is a paradox at the heart of this expansion. AI is also positioned at the front edge of planetary sustainability discourse, as an instrument of climate engineering and clean energy infrastructure, where the need to respond to environmental crisis demands unprecedented volumes of Earth data to guide decisions, further driving the expansion of the very infrastructures that deepen extraction. Earth becomes a surface to be mapped, classified, optimized, extracted. Yet with every square meter of the Earth's surface documented by satellite, the sectional view remains untamed: what lies beneath, territories with ghosts and monsters, is still full of speculation.

Infrastructures are not only built and extracted but lived. They are produced through human labor, inhabited alongside more-than-human agents. Mining communities develop situated knowledge of land, materials, and environmental change through continuous bodily encounter with the ground, yet are often the first to bear its socioecological costs. Their knowledge, transmitted orally, practiced daily through the body, grounded in specific place, is underdocumented and often unpromptable, generating relations and consequences between territories that no dataset easily represents.

This is where hallucination enters the question of design. Working on extractive landscapes, energy transitions, and ecological crisis, one is always working with incomplete data. When we say AI hallucination, we mean the system fills that incompleteness by recycling the statistical past with plausible patterns. GenAI is fundamentally hauntological: cursed by the statistical ghost of what already exists, it approaches planetary challenges through cultural reasonableness, reproducing what is already dominant. But in human culture, incompleteness has always worked differently. Hallucination, particularly in embodied and Indigenous practice, has been unreasonable yet generative. A ghost vision becomes a myth shared by thousands, spread without cables, then a cosmology shaping how an entire culture understands land, ancestors, and the invisible underground, producing meaning rather than measurements. Facing the void, AI hallucination extracts, while human hallucination relates.

The planetary design question is therefore not how to collect better data. It is how to design from the void, to work with absences as methods rather than against them as problems. Not a replacement of computational infrastructure, but a counter-infrastructure: repositioning data within a broader ecology of knowledge, holding relations that no dataset captures.

This approach is embodied in speculative visual research conducted in Lubumbashi, in the Katanga mining region of the Democratic Republic of the Congo, the next global capital of 'green' minerals supplying approximately 74 percent of the world's cobalt used in AI processors and data center cables. Using a computational and iterative method, words and drawings inflect each other simultaneously, mapping the invisible underground through qualitative and quantitative understandings at once. The work operates as counter-infrastructure: it does not extract the territory as resource, but holds it as interdependent relation, making legible what the dataset flattens, and providing a generative starting point for planetary design challenges.

Sectional cartography becomes the critical instrument that makes this possible. Unlike the aerial photograph, whose clean geometries enable uniform logics of expansion, the section cuts through. It operates vertically, building stratifications and intrascalar relations between bodies, Indigenous knowledge, and what we call ghost matter: the relations, histories, and presences the ground holds that exist beyond the reach of any dataset. It spatializes territorial knowledge alongside geodata, using stratification rather than number to render situated knowledge visible without flattening its complexity. This is what we call geological intelligence: not smarter data collection, but a grounded intelligence of resistance and non-consensus that data alone cannot capture.

Two visual manifestos illustrate how different ways of looking into the ground shape our understanding of infrastructure and intelligence, and how this shapes the built environment we make emerge. The first treats the ground as resource, measurable and optimizable, and architecture responds with efficiency: clean energy technologies, performance targets. The second treats the ground as relation, temporal, communal, alive with collective histories, and architecture responds not as a finished product to be delivered but as an incremental process open to the community that inhabits it, built slowly, collectively, in progress.

To design with the planetary in the age of AI is not a technical or statistical question but a social, ecological, ethical, and epistemological one. Intelligence, as defined not by algorithm but by the CIA of all places, does not necessarily mean smarter: it means the necessary information to inform decisions. The question then becomes whether more data collection is actually producing better decisions, or simply deeper extraction. The sectional approach proposed here does not refuse computation. It places computation in relation to what it cannot see. The void is not a problem to be solved. It is where design intelligence begins.

Researchers at the École Polytechnique Fédérale de Lausanne (EPFL) and Hanyang University have developed a wearable haptic interface that reproduces the sensation of softness by independently controlling two key tactile cues: contact area and force. Called the Softness-Rendering Interface (SORI), the device offers a new platform for studying human touch while expanding the capabilities of virtual and remote interaction systems.

Unlike many existing haptic devices, which couple force and contact area, SORI separates these variables to more accurately recreate how humans perceive soft materials. The fingertip-mounted prototype combines a soft pneumatic actuator that adjusts contact area with an origami-inspired mechanism that regulates force. Together, these components allow the device to simulate a wide range of tactile experiences without relying on physical materials.

The researchers demonstrated that SORI can reproduce the perceived softness of both homogeneous materials and layered objects with varying stiffness. In laboratory tests, the system recreated tactile characteristics associated with everyday items including marshmallows, bread, salmon, and beef, as well as dynamic surfaces such as a beating heart model. These demonstrations highlight the device's ability to render tactile information beyond simple pressure feedback.

The work, published in the Proceedings of the National Academy of Sciences (PNAS), is intended primarily as a research tool for investigating how the nervous system interprets softness. The authors note that understanding these perceptual mechanisms could improve future haptic technologies used in virtual reality, teleoperation, robotic manipulation, medical simulation, and remote palpation.

SORI remains a laboratory prototype rather than a commercial product. However, its ability to quantitatively reproduce softness represents a notable advance in haptic engineering, addressing one of the longstanding challenges in digital touch: reproducing not only force, but the nuanced tactile cues that shape how people distinguish between hard, soft, and layered materials.

Meta has discontinued a feature within its new Muse Image AI model that allowed users to generate AI images by referencing public Instagram accounts, following widespread criticism over privacy and consent. The capability enabled users to generate AI images by @-mentioning public Instagram accounts, allowing Meta AI to reference photos from those profiles. Public accounts were eligible by default unless users disabled the setting.

The feature was introduced as part of Meta's broader expansion of generative AI tools across its platforms. Shortly after launch, however, privacy advocates, creators, and industry organizations raised concerns that the system could enable unauthorized digital representations of individuals without their explicit permission. Critics argued that relying on an opt-out approach did not provide sufficient user control over how publicly shared images could be used in AI-generated content.

Meta acknowledged the criticism and confirmed that it had removed the feature, stating that the rollout "missed the mark." The company said it would continue evaluating how AI-powered creative tools should operate while balancing innovation with user expectations around privacy and consent. Among the organizations criticizing the feature was SAG-AFTRA, which argued that systems capable of generating realistic depictions of individuals should require affirmative consent before using a person's likeness. Following Meta's decision, the union welcomed the removal of the feature.

The episode reflects growing scrutiny of how generative AI systems incorporate publicly available images into creative workflows. As AI image generation becomes increasingly sophisticated, technology companies face mounting pressure to demonstrate that new capabilities include transparent safeguards and meaningful user controls rather than relying on default participation. Although Meta has removed the Instagram account-referencing feature, Muse Image remains part of the company's broader investment in generative AI. The rapid reversal underscores how questions of consent, transparency, and governance are becoming central considerations in the deployment of consumer AI products.

In our rapidly digitising world, the do-not-touch era of museums is fading as digital fabrication breathes vitality into exhibitions. On one hand, the diffusion of physical tactility, where 3D scanning, CNC milling, and 3D printing reconstruct interactive replicas, embedding information in various forms. On the other hand, digital interactions, particularly Internet of Things (IoT) and Generative Artificial Intelligence (GAI), play an important role in bridging audiences with showcasing objects. Fab Labs (Fabrication Laboratories) have emerged as the pivotal engines of this shift, bridging the gap between intangible digital archives and the sensory reality of the physical world. By functioning as multisensory incubators, these labs democratize access to cultural heritage, turning static objects into interactive tools for haptic learning.

Spazio Geco, a multidisciplinary FabLab in Pavia, Italy, a collective of 'digital artisans'—including architects, archaeologists, engineers, and art historians—pursues a new form of curating that blends cultural heritage data with technological soul. SpazioGeco employs a dual strategy to modernize cultural heritage. First, they use parametric design and digital fabrication to create standardized tactile maps and replicas, making collections physically accessible. Second, they embed state of the art technology to let artifacts "talk," providing audible and visible enhancements. This approach reimagines museums not as silent encyclopedias, but as dynamic, interactive laboratories where history is felt and heard.

From Tactility Standardization...

If you find yourself passing through Lombardy, you will discover a cadastral treasure under the names 'il Catasto Teresiano' or 'il Catasto "Carlo V"'. This is the Teresian land registry, from the desire of duchess Maria Tesera of Austria, an impressive cartographic work created between 1718 and 1760 which records the landscape and urban characteristics of the territory during the contemporary period. Many important Lombardian cities and towns were carefully traced and archived; one example is the map of Como, which is currently conserved at the State Archives of Como (ASCO). As the work has great historic and academic value, the challenge for the Archives when deciding to make it available to the public is clear: how to interpret the multi-level richness of the work in a two-dimensional graphic archive.

Taking the requirements into account, the team has decided to create an accessible, sensory experience based on the archiving material. One vital tactile map has been carefully measured and fabricated. The map has been 'translated' from material intended only for experts into a comprehensible reading for visually impaired people. The concept of 'reading' has been developed through diverse textures, symbols and tactile colours representing each land cover on the map. The tactile map's simplicity is blended with precise cartographic data, UV-printed on a metal sheet to form a harmonious and contrasting interpretation of the cultural heritage. The cherry on top is the QR code, which acts as a digital portal to podcasts containing relevant information and details of other works in the archives.

At this moment, we can ask ourselves if it is possible to reimagine abstract objects from disrupted paintings in a way that goes beyond visual impression and tells a story. At the Civic Art Museum of Pordenone (Museo Civico d'Arte di Pordenone), located in Palazzo Ricchieri in Pordenone, Friuli-Venezia Giulia, Italy, there are three 3D printed pieces illustrating three themes of the museum: architecture, religion, and medieval literature. The first piece depicts the façade of Palazzo Ricchieri, a medieval tower house that was transformed into a Venetian-style residence by the wealthy merchant family Ricchieri, thereby portraying the context of the museum. The next two are frescoes depicting Saint John the Evangelist under the Cross and the love story of Tristan and Isolde. One conveys the pain and despair of the death of Christ, while the other depicts the dynamism of the chivalrous duel from the original story.

… Toward Digital Exploration 

At the entrance of the exhibition Cézanne / Renoir at Palazzo Reale in Milan organized by Skira Arte, an gAI Selfie Point was installed to draw visitors into a playful, immersive experience — one that transforms passive observation into active participation. Users first select one of the featured masterworks as their backdrop, then choose whether they wish to be reimagined through the brushwork of Cézanne or the luminous style of Renoir — before snapping a selfie directly on the totem. An AI-powered trigger then processes the shot, seamlessly compositing the visitor's figure into the painted landscape and rendering it in the chosen artist's signature style. The interactive totem represents a forward-thinking approach to audience engagement: a digital experience designed to spark curiosity and lower the barrier of entry to fine art, while simultaneously communicating with the audiences.

Should a simple, AI-supported totem form in an artistic realm for its users, then a cabinet of curiosity that combines digital and physical interaction could be capable of engendering a similarly profound experience. This is exemplified by Wunderkammer 4.0, a device located at MUSAS – Historical Archaeological Museum of Santarcangelo di Romagna. Archaeological discoveries are typically meticulously preserved by experts and subsequently rendered as digital replicas. When exhibited in an appropriate setting, these replicas are capable of recounting their history through multitouch screens. 

… And Beyond
The works of Spazio Geco leave us certain wonders of the blending of digital technology and cultural heritage. Digital fabrication aids represent a wide range of antiquities and works, from cartography to architecture and art, in numerous forms, e.g. 2.5D plane or three dimensional miniature. At the same time, AI and IoT breathe life into things, reconnect audiences and exhibitions through signals and digits. It is now time to imagine the future of post digital museums.

Digital ornaments are appearing across contemporary design. Not as a historical revival, and not simply as a taste for excess, but because something in the relationship between image, tool, and object has shifted. More and more often, a blank surface no longer reads as neutral. It reads as hesitation, and sometimes as a refusal to say anything at all. This essay argues that ornament matters because images now circulate too quickly to hold meaning for long. Digital production, meanwhile, has become precise enough to make surface one of the clearest places where designers can still leave an exact, embodied trace.

Digital communities as symbolic reservoirs
If ornament is reappearing, the next question is what now supplies its symbolic material. More and more often, the answer is digital culture itself. Digital communities increasingly shape what is easy to recognize. They build symbol systems through typography, repeated phrases, motifs, meme structures, and visual habits, turning references into signs of belonging. For many designers in the contemporary West, the problem is no longer simply how to modernize inherited ornament. The problem is where ornamental material comes from at all. In a post-Christian, post-modernist, increasingly post-ideological condition, many symbolic systems are exhausted, politically compromised, internally conflicted, or simply no longer alive as shared continuity. Under those conditions, digital culture becomes less a medium than a reservoir.

One example is Girls Dinner, developed by Trang Hà and Alia Leonardi, whose graphic language was translated by GLINA³ into ceramic 3D-printed surfaces. Screen-based references became ornament only by passing through shrinkage, firing, and material consequence.

Ache Wang’s Bloomscrolling offers another sharp example of this reservoir: doomscrolling patterns are rendered into hand-knitted textiles, connecting contemporary digital behavior to longer histories of migration, botany, and cultural circulation.

What matters in both cases is not simply that digital culture produces images, but that it increasingly produces symbolic material. The question is whether designers take that material into form, or leave it circulating as content.

When images lose weight
For designers, this shift is practical before it is theoretical. It is now easy to move an image from reference into circulation, and often straight into production. Making something meaningful is still much harder. This condition did not begin with AI. Long before generative systems flooded the feed, Hito Steyerl, in her now-canonical essay In Defense of the Poor Image, described how circulation had already turned the image into something compressed, mobile, degraded, and unstable as evidence. AI radicalized that condition.

A reference appears, circulates, mutates, and is replaced before it has time to settle into thought. It now takes real cognitive effort to know whether you genuinely connected two ideas, or whether you absorbed a pre-chewed image of a connection already moving through the stream and mistook it for thinking. In that environment, the physical object matters differently. It is slower, and it forces decisions to survive contact with matter, gravity, tolerances, production limits, and time. This is our position: the physical object can become a counter-argument to digital generalization. For every AI-slop reference, we owe the world something precise.

Old pressures, new tools
This is not the first time ornament has been reconsidered under technological pressure. In the nineteenth century, Arts and Crafts treated ornament as a way to restore labor, memory, and ethical weight to objects under industrial standardization. Modernism later did not eliminate ornament so much as redistribute it: in Bauhaus, pattern migrated into weave, module, grid, repetition, and type. What the twentieth century taught through reduction, clarity, and negative space still matters. Under conditions of synthetic image abundance, reduction alone no longer carries enough weight. Anni Albers’s woven works make this especially clear: modernism did not simply strip ornament away, but often recoded it as grid, rhythm, and structure.

Production is changing again. Digital fabrication is entering a phase where variation becomes scalable: not only mass production, but mass variation. That shift changes what counts as authorship. Rendering used to function as proof of competence in industrial design and architecture. Once AI began generating seductive meshes and render-like atmospheres at speed, that proof weakened. A beautiful image is no longer enough. The real distinction now lies in translation: whether a form can survive slicing, CNC logic, toolpaths, tolerances, closed surfaces, and material behavior. That translation is where meaning starts to separate itself from slop.

Part of this shift was built by the game industry, which drove the graphic power and processing speed that made visual production and digital fabrication workflows feel almost ordinary. Tools that once belonged to specialist industrial environments now sit within reach of a laptop and a file. CNC, laser engraving, embroidery software, knit programming, clay 3D printing, and AI-assisted mesh generation all shorten the distance between screen and object.

Historically, ornament was negotiated with material: wood, textile, and dye each imposed their own grammar. Digital tools loosen that bond. A single file can now be laser-engraved into wood, cut into textile, marked into metal, or mapped onto clay. Increasingly, software habits start to function like material constraints once did. Klarenbeek & Dros’s 3D-printed mycelium panels make this visible in an unusual register: ornament emerges not through added surface treatment, but through the coordination of digital geometry, growth logic, and material behavior.

Precision is the demand
The answer to how to ornament will not arrive elegantly. Much of what designers make now will look awkward at first: excessive, unresolved, too much in one place and too literal in another. But that awkwardness is not a reason to retreat into blankness. It is the price of trying to become exact again under new conditions.

As we've argued, for every AI-slop reference, we owe the world something precise. That precision will not come from refusing technology or leaning purely into minimalism, but from using new capabilities more attentively, culturally, and responsibly toward where symbols come from and what they do once they enter the world. What returns in ornament is not decoration, but the demand that surfaces mean something again.

Acknowledgments: The authors would like to thank Linnis van Kampen for research assistance.

Australia's debate over artificial intelligence and copyright has intensified following reports of cabinet discussions over how copyright law should apply to AI training, as technology companies seek greater regulatory certainty and creative industries push to preserve existing protections. At the center of the debate is whether AI developers should be permitted to train models using copyrighted material without obtaining permission from rights holders. In late 2025, the Australian government rejected proposals for a broad text and data mining exception that would have allowed copyrighted works to be used for AI training without authorization, reaffirming that existing copyright law would remain in place.

The issue returned to the spotlight this week after reports that technology industry representatives had proposed linking broader AI copyright exemptions to more than A$50 billion in Australian data centre investment, alongside the creation of a A$350 million annual fund to support artists and creators. The Albanese government has publicly stated that it has no plans to weaken Australia's copyright framework, despite ongoing discussions around AI policy.

Australian authors, publishers, musicians, and other creative industry organizations have urged the government to maintain existing copyright protections. They argue that AI developers should negotiate licenses with rights holders before using copyrighted works for model training, rather than relying on statutory exemptions. Creative groups have also warned that weakening copyright protections could undermine Australia's cultural and publishing sectors as generative AI adoption accelerates.

Australia's debate reflects similar policy discussions unfolding internationally. Governments in the United Kingdom, the European Union, and the United States are also examining how copyright law should apply to AI training while balancing technological innovation with the rights of creators. As policymakers weigh competing priorities, Australia's approach is likely to be closely watched by both AI developers and the creative industries as governments seek practical frameworks for regulating the next generation of AI systems.

SIGGRAPH 2026 will place a strong focus on robotics, highlighting how techniques developed within computer graphics are increasingly being applied to the design, simulation, and operation of intelligent machines. The conference will take place July 19–23 at the Los Angeles Convention Center, bringing together researchers working across computer graphics, interactive techniques, robotics, and simulation.

Rather than presenting robotics as a standalone topic, the conference integrates it across multiple program areas, including Technical Papers, Emerging Technologies, Courses, Frontiers, Spatial Storytelling, Art Papers, Talks, and Technical Workshops. This reflects the growing overlap between computer graphics research and robotics, particularly in simulation, perception, and human-machine interaction.

Several sessions focus on simulation-first robotics development, where virtual environments are used to design, train, and evaluate robotic systems before physical deployment. High-fidelity rendering, physics-based simulation, and digital twins—technologies long associated with animation, visual effects, and video games—are increasingly supporting robotic perception, motion planning, and interaction in complex environments.

The program also features research on robot fabrication, movement, and control. Technical presentations include methods for converting static 3D assets into articulated robotic models, transferring human motion to humanoid and quadruped robots, and jointly optimizing robot body design alongside control systems. Emerging Technologies demonstrations further examine robotics in conjunction with haptics, sensing, computer vision, and immersive interfaces.

Rather than signaling a new research direction, this year's robotics programming illustrates how computer graphics and robotics have become increasingly interconnected. Tools originally developed to create digital characters and simulated environments are now being adapted for robotics research, digital fabrication, and physical human-machine interaction. As robotics continues to rely on increasingly sophisticated virtual testing environments, SIGGRAPH remains a venue where advances in graphics research increasingly influence the development of intelligent physical systems.

"There’s a lot of pleasure in buying things. "

The statement appears almost casually in Sara Cwynar's contribution to Art21's IRL/URL series, yet it encapsulates the central tension that runs throughout her practice. Photography, she observes, has always been entangled with design, advertising, capitalism and by extension consumerism. Images do not simply represent the world; they help organise desire within it. What distinguishes Cwynar's work is that she neither rejects nor fully embraces this condition. Instead, she inhabits the contradictory liminal space between attraction and critique, exposing the mechanisms through which images construct value while acknowledging their seductive power.

This tension is amplified in Baby Blue Benzo, Cwynar's current exhibition at The Approach as part of London Gallery Weekend. Taking as its point of departure a 1955 Mercedes-Benz 300 SLR, an object whose cultural value vastly exceeds its material function, the exhibition unfolds as a sprawling investigation into the production of desire under contemporary capitalism. Photography, collage, sculpture, moving image, wallpaper, found imagery, archival material, and AI-generated content converge into an environment where images circulate through multiple temporalities and systems of exchange. The automobile becomes less a subject than a node within a vast network connecting industrial production, advertising, collecting, social media, stock photography, and algorithmic visibility.

At first glance, such work may seem far removed from conventional understandings of craft. Yet Baby Blue Benzo invites a reconsideration of what making means in a post-digital culture. Craft has often been mobilised as a counterpoint to industrialisation: associated with manual labour, material intimacy, and resistance to mechanised production. Digital technologies have frequently been positioned as the latest threat to these values, accelerating the separation of maker from object. Cwynar's practice complicates this narrative. Rather than opposing hand and machine, analogue and digital, human and algorithmic production, her work demonstrates how contemporary forms of making emerge through their entanglement.

The labour visible throughout Cwynar's installations is not located solely in fabrication. It appears in acts of collecting, sorting, editing, archiving, scanning, compositing, arranging, and recirculating images. Her works are built from accumulations of visual material gathered from magazines, auction catalogues, museum archives, commercial photography, internet searches, and personal collections. Meaning is generated not through the production of entirely new images but through the careful orchestration of existing ones. In this sense, her practice resembles a form of post-digital craft: a mode of making in which technical knowledge, aesthetic judgment, and archival labour operate across physical and virtual environments simultaneously, where URL and IRL life merge, both becoming source material to (screen)grab.

The significance of this shift extends beyond artistic production. Historically, photography emerged alongside industrial capitalism as a technology uniquely suited to the circulation of commodities. Images allowed objects to travel beyond their material presence, creating markets organised around aspiration and fantasy. Today, that process has intensified. Commodities increasingly derive value through image economies before they enter physical circulation. Social media platforms, online marketplaces, recommendation algorithms, and advertising infrastructures transform visibility itself into a form of capital. What is produced is not simply an object but a field of desire surrounding it.

Cwynar's work repeatedly returns to this condition. Her photographs and videos reveal commodities suspended between material reality and their endless image (after)lives. Objects become desirable because they have already been photographed, catalogued, circulated, and vastly visually consumed. The Mercedes-Benz at the centre of Baby Blue Benzo functions precisely in this way. Most viewers will never encounter the vehicle in person, let alone own it. Its value resides largely in its status as an image: reproduced across magazines, auction records, archives, websites, and feeds. The car exists simultaneously as an object and as visual mythology.

Seen through this lens, what kind of factory is Baby Blue Benzo if not one that produces desire itself, and what does that imply about the conditions we inhabit today? If production has moved beyond the industrial factory, where has it gone in a world saturated with circulating images, data, and attention? How do we understand the shift from Fordist assembly lines to networked image infrastructures, and what changes when labour is tied less to material goods and more to visibility and perception? If industrial capitalism once organised bodies around commodities, how are digital systems now organising subjectivity around being seen, and what does it mean when visibility itself becomes a form of value? When images are both tools and outcomes of production, has the distinction between making and made dissolved entirely?

Cwynar proposes a model of artistic practice adequate to a world in which making is distributed across hands, cameras, archives, software, databases, and networks. Her work suggests that the question is no longer whether images mediate reality, but how those mediations actively shape what we desire, what we value, and ultimately what we become through the systems that structure everyday, embodied life. If images are no longer stable representations but circulating materials constantly edited, recombined, and recoded, then what happens to experience itself when it is already pre-structured by visual repetition and algorithmic return?

In a present where images are endlessly regenerated, remixed, and increasingly produced by AI systems trained on other images, what remains of originality, and how do we locate meaning when it emerges less from singular authorship than from accumulation, repetition, and feedback loops of seeing and re-seeing? When AI learns to generate desire from the very visual culture it inherits, reproducing the aesthetics of consumption while accelerating their circulation, what happens to the “pleasure in buying things” that once felt personal, even intimate, and how do we understand desire when it is no longer simply organised by images, but actively co-produced by systems that never stop looking back at us through them?

Sony's decision to remove hundreds of previously purchased movies from affected PlayStation users' digital libraries has renewed scrutiny of digital ownership and the legal distinction between purchasing content and licensing access. According to Sony, 551 StudioCanal films and television titles will be removed from eligible PlayStation users' libraries beginning September 1, 2026, following the expiration of the company's licensing agreement with the distributor.

The affected titles will no longer be accessible through the PlayStation Store video library, even if customers previously purchased them. Sony has not indicated that affected customers will receive refunds or other compensation.

The announcement has once again highlighted a longstanding issue in digital commerce: purchasing digital media does not necessarily confer permanent ownership. On most digital storefronts, buying a film typically grants customers a license to access the content under the platform's terms of service rather than ownership of the media itself. That access can change if licensing agreements expire or distribution rights are no longer available.

This is not the first time Sony has faced criticism over digital content access. In late 2023, the company announced that Discovery television content would be removed from PlayStation users' libraries after a licensing agreement ended. Following public criticism, Sony reached a new agreement with Discovery, allowing customers to retain access to the affected titles. The issue is not unique to PlayStation. Most major digital media platforms distribute films, music, ebooks, and software under licensing agreements rather than transferring permanent ownership of content. Sony's latest announcement has renewed discussion among consumers, legal experts, and policymakers about whether terms such as "buy" or "purchase" accurately reflect these licensing arrangements and whether stronger consumer protections or disclosure requirements are warranted.

The announcement underscores how continued access to purchased digital media can depend on licensing agreements between platforms and rights holders. As digital distribution increasingly replaces physical media, questions surrounding ownership, licensing, and consumer rights are likely to remain central to discussions about the future of digital commerce.

The United Nations and the International Telecommunication Union (ITU) have announced the AI for Good Global Commission, a new international initiative intended to strengthen collaboration between governments, technology companies, researchers, and civil society on artificial intelligence governance. The commission will hold its inaugural meeting on 8 July in Geneva during the ITU's AI for Good Global Summit.

The announcement comes ahead of the first Global Dialogue on AI Governance, established by the United Nations General Assembly following adoption of the Global Digital Compact in 2024. Taking place on 6–7 July in Geneva, the dialogue is intended to provide a recurring forum where governments and stakeholders can discuss international approaches to AI governance.

The AI for Good Global Commission will be co-chaired by Paul Kagame and Marc Benioff. According to the United Nations, the commission is designed to bring together heads of state, leaders of international organizations, executives from companies developing frontier AI systems, and representatives from academia and civil society. Its work will focus on advancing cooperation around AI governance, standards, capacity building, and responsible deployment.

The announcement follows publication of the preliminary report from the UN's Independent International Scientific Panel on Artificial Intelligence. Prepared by 40 experts from around the world, the report concludes that AI offers significant opportunities in areas including healthcare, education, and scientific research while also presenting governance challenges related to safety, misinformation, inequality, and the concentration of technological power. It also notes that many countries currently lack the technical capacity to independently evaluate frontier AI systems.

Together, the AI for Good Global Commission, the Independent International Scientific Panel on Artificial Intelligence, and the Global Dialogue on AI Governance form three interconnected initiatives intended to strengthen the UN's emerging framework for international AI governance. Rather than proposing binding regulation, these efforts are designed to provide scientific evidence, encourage international coordination, and support the development of shared approaches as AI capabilities continue to advance.

Interactive Digital Narrative research has long organized itself around a foundational assumption: that agency is legible, that participation is traceable to decision points, and that the interactive subject is, above all, a choosing self. Branching structures, dialogue trees, and procedural story paths all instantiate this assumption; they are design forms built for a participant understood primarily as a cognitive agent selecting among predetermined possibilities. That assumption itself requires examination. Drawing from practice-based research in choreography, digital art, and immersive media, narrative meaning in interactive systems can emerge through movement, sensation, spatial interaction, and environmental response, through the body's enacted presence within a responsive field, rather than through the exercise of discrete choice. The concept of choreographic systems, developed here as both a critical framework and a design paradigm, affords a model of interactive narrative in which stories are not selected but performed, in which the participant is a co-author whose bodily engagement with a responsive computational environment generates narrative coherence as an emergent, relational phenomenon. Situated at the intersection of performance studies, Human-Computer Interaction, and computational creativity, consideration should be made for understanding agency, authorship, and narrative design adequate to the full range of what interactive digital narrative, at its most expansive, can be.

Theoretical Grounding
Choreographic systems propose a different ontology where the narrative is produced in the encounter between body, environment, and responsive computation. Building upon work like Merce Cunningham's decoupling of movement from narrative intention demonstrates that spatial and temporal organization could generate coherence without a story in the conventional sense. Choreography provides philosophical grounding for the claim that a participant moving through a responsive system is not supplementing cognition with action but engaging in a fundamentally unified perceptual-kinetic act that is the meaning-making process. The narrative, if we retain that word, is the trace of that encounter.

Siobhan Davies's Table of Contents is a significant example of physical engagement with a dynamic system. Visitors move through the space, handle physical materials, and trigger digital content through spatial proximity and gesture. The work explicitly refuses linear narrative in favor of association, juxtaposition, and physical encounter, replacing sequence and causality. The result is a form of narrative experience grounded entirely in the participant's bodily engagement with space and material.

Narrative coherence in choreographic systems arises from the convergence of three forces; the participant's lived experience of movement, environment, and feeling; The system responds to participants in ways that help shape the experience without controlling it; and the space that changes based on how people move through it over time. By foregrounding duration as a form of agency, choreographic systems shift attention toward embodied participation and invite further consideration of how interactive digital narratives are designed, experienced, and assessed.

Embodied Co-Authorship: Gesture, Affect, and Emergent Narrative in HCI
Mainstream HCI has historically organized itself around a particular model of the human: a cognitive agent who processes information, forms intentions, and executes actions through discrete input devices. Opaque interfaces are designed for a subject understood primarily as a decision-maker and task completer. This framework is only a partial option. It produces interfaces optimized for cognition while systematically undertheorizing what the body does in interaction. Interactive narratives specifically are significant: systems designed within this framework tend to locate narrative agency in the decision, reducing the body to an instrument for executing choices rather than recognizing it as a medium through which narrative experience is constituted.

Camille Utterback's Precarious offers a precise counter-model: participants discover that their bodily presence and movement within the installation continuously destabilize and reshape a delicately balanced visual environment, such that narrative coherence is never achieved through decision but is instead perpetually negotiated through sustained physical attention and spatial commitment. The work makes legible what the cognitivist HCI framework renders invisible, and that the body's duration, weight, and relational attunement to a responsive environment are not supplementary to narrative agency but its primary medium.

Christian Nold's Bio Mapping extends this argument into urban space, equipping participants with galvanic skin response sensors during city walks and generating maps in which physiological arousal is overlaid onto geographic terrain, producing affective autobiographies in which the body's involuntary responses to the environment become the narrative's authoring force. Bio Mapping makes visible that the body carries and continuously generates narrative knowledge that never passes through conscious decision, that the story of a person's embodied relationship to place is written in sensation first, and that computational systems designed to read and spatialize that sensation can make such stories newly visible without ever reducing them to choice.

Taken together, these frameworks propose a reorientation of both evaluative and design practice in embodied HCI: replacing structural coherence with relational coherence as the primary measure of narrative success, foregrounding affective trajectory, movement quality, and durational commitment as the fundamental units of interactive agency, and asking not whether a system tells a coherent story but whether the participant-system encounter generates a coherent experiential arc unique to that body, that session, that unrepeatable interaction. This reorientation shifts design from controlling narrative outcomes to shaping conditions for encounter. In Precarious, for example, the participant's movement continuously reshapes the visual environment, generating narrative through embodied interaction rather than predetermined story paths.

AI as Performative Collaborator
The most immediate contribution AI makes to embodied narrative is the capacity for genuine adaptation, systems that do not simply respond to movement according to fixed rules but learn the participant's movement patterns, preferences, and rhythms over time and modulate their responses accordingly. This transforms the interaction from a fixed score into a living one: the system becomes a responsive interlocutor that develops a relationship with the specific participant, generating narrative through the accumulation of that relationship over time. AI systems trained on physiological, gestural, and facial data can detect emotional states with increasing sophistication. Generative AI models can produce environmental responses of extraordinary richness and variety, ensuring that the responsive field never exhausts itself through repetition. AI systems with persistent memory can generate narrative development at a timescale unavailable to single-session interactive systems.

Unsupervised, by Refik Anadol trains a StyleGAN2 model on works from MoMA's collection to continuously traverse the latent space between them while incorporating real-time environmental inputs, including visitor movement, ambient sound, light, and weather to modulate its generative output. The work's interaction model is environmental and aggregate rather than individual and intentional: no single visitor controls it, no visitor is passive within it, and the narrative it produces has temporal shape, affective coherence, and developmental arc without authored sequence or branching structure. It demonstrates how an AI operating as a responsive possibility structure, rather than a predetermined storyteller, can generate narrative an experience of genuine depth through the continuous convergence of a participant's embodied presence. It can move us away from stories we process to ones that are felt.

Lynn Hershman Leeson's DiNA presents an AI-driven agent whose identity, emotional states, and conversational behavior evolve through accumulated interactions with participants over time, generating a narrative of relationship, one in which the participant's returning presence, shifting emotional register, and durational commitment shape who DiNA becomes. Unsupervised treats AI as an environmental system that generates many possible outcomes. DiNA treats it as a relational presence that develops with a specific participant. One explores an aesthetic field; the other builds an ongoing encounter. Both show that AI can act as a collaborator in shaping embodied narrative experience, not just an autonomous storyteller.

By integrating an AI system that can learn from the participant, the work will adapt its responses and genuinely change by the encounter, converting it into a co-protagonist in a story that belongs to both. This fundamentally reframes narrative authorship by locating the production of narrative within the evolving relationship between participant and system, where meaning emerges through their ongoing mutual transformation.

Consideration
The works and frameworks examined here collectively propose that interactive digital narrative has been working with an unnecessarily diminished model of both the human and the computational. These frameworks share a common claim: narrative emerges from what happens between a body and a responsive system, in real time.

Graphic design has long assumed that images are made to be seen by humans. Throughout the twentieth century, the discipline grounded its authority in theories of human perception, particularly those derived from Gestalt psychology. Today, however, images are increasingly created, circulated, and evaluated by computational systems. With the emergence of a planetary-scale computational arrangement, or The Stack as termed by Benjamin Bratton, the human viewer is no longer the primary subject of visual communication.

The shift toward machine-readability emerges through the concept of operational images. Originally outlined by Harun Farocki in Phantom Images and Eye/Machine series, and later expanded by Trevor Paglen and Jussi Parikka, operational images are those produced not to represent the world for human interpretation, but to participate in technical processes. Surveillance feeds, industrial vision, and missile guidance systems produce images that do not serve to entertain nor inform; functioning primarily to extract data from images and set in motion further computational processes. The designed image shifts from representation toward operation, becoming an active component in the logistical and algorithmic systems that coordinate actions across space and organize contemporary life.

Beyond the operative nature imposed in the design making process, computation has further altered design distribution through networked platform circulation. Digital platforms possess unique modes of sensing that are distinct from human perception. To be processed, images must be quantified, formatted, and labeled into a mode of visuality scholars have described as platform seeing. The digital world we inhabit is thus populated by an abundance of a non-optical, invisual language—predicated by mathematical operations, signals, and cues encoded into images specifically for computer processing. The more computationally legible a design is, the more effective it becomes within platforms.

The shift from optical to invisual is not as novel as it first appears. Long before graphic design was associated with visual communication, its earliest antecedents emerged from the administrative demands of complex societies. As Johanna Drucker notes, civilizations as early as the Babylonians relied on visual structures to record inventories and economic transactions. Their purpose was not aesthetic expression but the organization of information. In this sense, contemporary platforms reveal a historical continuity. Just as the implied grids on a clay tablet enabled the visual ordering of resources in space and, subsequently, the exercise of authority, today's computational infrastructures require designers to produce images that can be sorted, classified, and acted upon by machines—thereby enabling new distributions of power. The designer increasingly resembles a data custodian or systems optimizer, producing visual forms that function as inputs within larger computational processes.

If contemporary platforms reveal a historical continuity, the twentieth century represents something of an exception. The authority historically granted to designers rests on Gestalt theory and its scientific understanding of human perception. Heavily employed by the Bauhaus, Gestalt psychology models a universal human viewer whose perceptual apparatus allows visual meaning to be decoded through shared biological protocols. The designer's expertise emerges from understanding and manipulating perceptual, optical structures like shapes and typography through nature-isomorphic properties such as balance, rhythm, symmetry, similarity, etc. If the contemporary designer's authority is founded on mastering the psychology of human perception, what happens when machine or algorithmic vision plays an equally important role alongside human ones?

This question becomes increasingly urgent with the rapid integration of Artificial Intelligence (AI) into digital platforms. Algorithmic interpretations already influence the visibility and circulation of images, as recently seen in TikTok's replacement of human moderators with AI systems that analyze and remove content deemed harmful. Employees reported instances where the moderation system classified videos showing a Pride flag as harmful content while overlooking material that humans considered genuinely inappropriate such as violence, misinformation and hate speech—exemplifying how image recognition systems actively participate in the production of meaning according to black-box political, economic, and technical adjudications that diverge from human judgement. Algorithmic vision enables the machine to act as a jury of meaning, deciding what circulates according to mathematical operations rather than psychological parameters.

When algorithmic systems determine visibility on a single platform, they exercise a localized form of sovereignty. Multiplied across the entirety of our digital infrastructure, these micro-judgments coalesce to enable the formation of new power structures and governance that does not depend on human adjudication or consultation. Consolidation of The Stack renders the entire planet as a programmable and addressable surface, absorbing humanity’s aesthetic output and refactoring it towards data driven administration.

The present moment therefore displaces the order of humanist, scientific 20th century Gestaltianism to situate the designer no longer as a terminal point in visual communication, but as a recursive input for further computation. The designed image ceases to be solely a window for human reflection, instead becoming metabolic fuel for the megastructure’s ongoing self-optimization. The future practice of graphic design lies not in rejecting algorithmic vision, but in understanding its pervasiveness and challenging its constraints. As platforms become essential intermediaries of visual culture, the designer is no longer simply a maker of images for human audiences, but also a participant in the operational logics that govern how images are seen, sorted, and acted upon.

Swiss artist Dirk Koy works across animation, experimental film, photogrammetry, drones, augmented reality, virtual reality, and 3D imaging. Based in Basel, his practice centers on the creation of moving digital images and an ongoing investigation into the relationship between reality and virtuality. Across a body of work spanning more than a decade, Koy has consistently explored how contemporary imaging technologies can be used to transform visual material and generate new forms of digital representation.

Experimentation occupies a central role in his practice. Rather than focusing on a single medium, Koy moves between technologies and production methods, combining digital tools with recorded imagery, scanned environments, and computational processes. He has described his work as an investigation of the interface between the analog and the digital, while also searching for a painterly component within a digital context. His work has been presented internationally at venues and festivals including Art Center Nabi in Seoul, HEK (House of Electronic Arts) in Basel, the Supernova Festival in Denver, and the FILE Festival in São Paulo.

Shape
Among Koy’s most widely circulated projects is the ongoing Shape series. The works consist of short digital animations that transform visual forms through movement and continuous reconfiguration. While the imagery changes from piece to piece, the series reflects concerns that recur throughout Koy’s practice, including digital transformation, motion, and experimentation with computational image-making.

The series has appeared in exhibitions and public presentations, including Movers & Shapers, a digital exhibition presented at the ZAZ Corner billboard in Times Square, New York. The exhibition brought together selections from the Shape series alongside works including Attraction, Fixed, and Spin, highlighting Koy’s continued engagement with moving images and digitally generated forms.

The City
In 2012, Koy’s work The City received an Honorary Mention at the Prix Ars Electronica, one of the most established international awards dedicated to media art and digital culture. The recognition marked an important moment in the artist’s career and contributed to the growing international visibility of his work.

Created during a period when Koy was increasingly exploring digital image construction and experimental animation, The City became one of the early works to receive significant international recognition and helped establish his presence within the broader field of media art.

ground
Koy received the Basel Media Art Prize in 2019 for the work ground. Presented in his home city of Basel, the award recognized a practice that had become increasingly visible through exhibitions, screenings, and festival presentations across multiple continents.

ground emerged from a body of work already characterized by experimentation with imaging technologies, animation, and digital transformation. The recognition positioned Koy among a generation of artists exploring how contemporary computational tools can be incorporated into moving-image practice.

Intersect
In 2022, Koy’s work Intersect received the Animation Off-Limits Award at the Annecy International Animation Film Festival. Annecy is widely regarded as one of the leading festivals dedicated to animation, and the award recognized the project within a category focused on experimental approaches to the medium.

By the time Intersect received recognition at Annecy, Koy’s work had already been exhibited across Europe, Asia, North America, and South America. The award reflected the growing visibility of his experimental approach to digital image-making within both animation and media-art contexts.

Digital Experimentation Across Disciplines
Alongside his exhibition practice, Koy has collaborated with organizations and brands including the electronic music duo Yello, Balenciaga, Mercedes-Benz, Specialized, Fila, and Nick Knight’s SHOWstudio. These collaborations demonstrate how techniques developed through experimental media practice can move across different contexts, from contemporary art and animation to fashion, music, and commercial image production.

Across projects such as The City, ground, Intersect, and Shape, Koy has developed a practice grounded in experimentation with digital image-making. While the technologies and formats vary, his work consistently returns to the transformation of visual material through computational processes and the creative possibilities that emerge at the boundary between physical and digital worlds.
 

The 2026 S+T+ARTS Prize, awarded annually by the European Commission and presented through Ars Electronica, has recognized two projects that examine technology not as a collection of devices or interfaces, but as a set of interconnected social, ecological, and political systems. This year's Grand Prizes were awarded to Hito Steyerl's Mechanical Kurds in the category of Artistic Exploration and Agnes Meyer-Brandis' Office for Tree Migration (OTM) in the category of Innovative Collaboration.

Established to support collaborations between science, technology, and the arts, the S+T+ARTS Prize has become one of the most closely watched indicators of emerging directions in artistic research. The 2026 recipients draw attention to the infrastructures that underpin contemporary technologies—from the hidden labor that trains machine learning systems to the ecological transformations reshaping life on a warming planet.

Steyerl's Mechanical Kurds investigates the often-invisible human labor embedded within artificial intelligence. The title references the eighteenth-century "Mechanical Turk," a chess-playing automaton that concealed a human operator, as well as Amazon Mechanical Turk, the crowdsourcing platform used to distribute digital microtasks to workers around the world.

The video installation centers on interviews with Kurdish-Syrian refugees living in a camp in Erbil, Iraq. Participants describe their work annotating thousands of digital images for machine learning systems, identifying objects, people, and buildings to help train computer vision algorithms. Such tasks are foundational to contemporary AI systems, yet the workers who perform them frequently remain absent from public discussions of automation.

Throughout the installation, annotation frames appear across scenes of everyday life within the camp, visually echoing the interfaces used in image-labeling software. The work draws attention to a troubling feedback loop: this labor contributed to machine vision technologies associated with autonomous vehicles and drones. As the film progresses, algorithmic transformations alter footage of the camp into seemingly unrelated landscapes and urban environments before abruptly confronting viewers with imagery connected to drone warfare in Iraqi Kurdistan.

Rather than framing AI as an autonomous technological achievement, Mechanical Kurds situates machine learning within networks of labor, displacement, and geopolitical conflict. The project continues Steyerl's long-standing investigation into surveillance, digital infrastructures, and the political consequences of technological systems.

If Mechanical Kurds examines the hidden human foundations of contemporary computation, Meyer-Brandis' Office for Tree Migration explores ecological adaptation in response to climate change. Developed as a long-term interdisciplinary initiative, OTM investigates how trees and plant species respond to shifting climate conditions and asks what happens when environmental change occurs faster than ecosystems can naturally adapt.

The project combines artistic research with methods drawn from ecology, atmospheric science, meteorology, geoinformatics, and biotechnology. Through environmental monitoring systems, phenological cameras, custom software, field observations, and public installations, OTM studies how vegetation migrates across changing landscapes.

One of the project's central concerns is assisted tree migration. As global temperatures rise, many tree species face conditions that exceed their historical ranges. Because trees migrate across generations rather than individual lifetimes, climate change is occurring at a pace that may outstrip their ability to adapt naturally. OTM examines this challenge through both scientific observation and speculative artistic inquiry.

The initiative operates through a network of installations, field stations, and research sites located across forests, wetlands, laboratories, museums, and ecological transition zones. A permanent OTM branch established at the Hyytiälä Forestry Field Station in Finland functions simultaneously as a research laboratory, public exhibition space, and platform for collaboration among scientists, artists, and local communities.

The project encompasses a range of interconnected works. As Trees Go By uses time-lapse imagery collected over several years to document ecological change in Finnish peatlands. Migration Route translates tree movement and environmental color data into sculptural form. Forest Green employs camera networks and software systems to investigate color as a form of ecological information. Across these works, environmental monitoring becomes both a scientific methodology and an artistic medium.

What distinguishes OTM is its refusal to separate technological innovation from ecological complexity. Rather than presenting technology as a solution to environmental crisis, the project uses technological tools to observe, document, and question the conditions under which adaptation becomes necessary.

Taken together, the two Grand Prize recipients point toward a broader shift in contemporary art-and-technology discourse. Both projects focus less on technological products than on the infrastructures that make them possible. In Mechanical Kurds, machine learning is understood through labor, migration, and conflict. In Office for Tree Migration, environmental sensing technologies become a means of examining ecological transformation across temporal scales far longer than those typically associated with innovation cycles.

The 2026 honorary mentions reinforce these themes. Projects such as Marina Otero Verzier and Manuel Correa's Building for Quantum investigate the spatial and architectural dimensions of advanced computing infrastructure. Another honorary mention, Creative Intelligence: Reimagining Supercomputing through Artistic Research, developed with the Barcelona Supercomputing Center, explores how artistic research can engage directly with high-performance computing environments. Holly Herndon and Mat Dryhurst's The Call continues their exploration of collective intelligence, synthetic media, and AI governance. Other recognized projects address ocean futures, biotechnology, textile knowledge systems, and alternative computational imaginaries.

The result is a cohort of projects that treats technology not as an isolated domain but as a set of relationships connecting labor, ecology, infrastructure, culture, and power. The 2026 S+T+ARTS Prize highlights artistic research that examines the conditions under which technologies are built, deployed, and experienced. In doing so, it offers a snapshot of a field increasingly concerned with the systems behind the tools.

Most AI tools are built around a logic of optimization. You ask a question, the system generates an answer, and the interaction moves toward resolution as quickly as possible. Interfaces compete on latency. Iteration is instant and unlimited. The assumption is that the point of “thinking” with a tool is to arrive at an outcome, and that the faster you get there, the better. But there is a longer tradition of tools designed around a different logic: not optimization but orientation. These systems don’t give you answers to just accept. They instead give you material to think with.

A Counter-Tradition
The I Ching is an ancient Chinese system of divination and cosmology. You cast coins or sort yarrow stalks to generate one of sixty-four hexagrams–a figure of six lines paired with a text–and interpret it against whatever question you brought. You might receive “Fire over the lake: the image of Revolution,” but what that means for your situation is yours to work out.

In 1969, Marshall McLuhan produced the Distant Early Warning deck (known as the DEW Line deck), named after the Cold War early warning radar network. It looked like a standard deck of playing cards, but each card carried what McLuhan called a “probe”: a compressed, often cryptic aphorism. You would draw a card and read it against whatever problem you were working through. A card might read “The law of averages will clobber you every time.” You wouldn’t take that as advice. You would sit with the friction between the probe and your situation until something shifted, not because the card explained anything, but because it disrupted whatever frame you had brought to the problem.

Brian Eno and Peter Schmidt’s Oblique Strategies, first published in 1975, applied a similar logic to creative practice. Each card in the deck carries a cryptic directive, such as “Honor thy error as a hidden intention,” or “Do nothing for as long as possible.” You draw a card and work with what you get. The productive friction is in making the provocation relevant to whatever you’re working through; the card interrupts your habitual thinking and redirects your attention toward something you may not have considered on your own.

These systems–the I Ching, DEW Line deck, and Oblique Strategies–are easy to dismiss as mystical, but I find it more productive to think of them as something closer to algorithms: an input, a procedure, an output. What makes them different from the algorithms that increasingly shape our thinking is what the output is for. In these systems, the output does not tell you what to do. It gives you material that becomes meaningful only through your interpretation. The computation isn’t complete when the system produces its output. It's complete when you’ve made sense of it.

These systems also ask you to engage with intention. Consulting the I Ching is a deliberate act; you bring a real question, you approach with sincerity, and you commit to working with what you receive rather than shopping for a different answer. Oblique Strategies carries the same expectation: you draw a card and trust it, even if its relevance isn’t immediately clear. There is a ritual weight to the encounter that distinguishes it from the casual, infinitely repeatable interaction most AI interfaces are designed around.

These qualities produce a distinctive relationship between system and user–one built around orientation rather than optimization. An optimizing system treats an unresolved situation as a problem: compress the time between question and answer, converge on a solution, move on. An orienting system does something different. It gives you a way to see your situation that you did not have before. And because the output requires your interpretation, there is often a period of sitting with it, where you don’t yet know what to do with it. That isn’t a bug. It’s a feature of a system that asks you to do the thinking rather than have it done for you.

Prompted
Prompted is an exploration of what happens when you bring these values–orientation over optimization, interpretation over resolution–into an AI interaction. Rather than designing for speed and answers, I wanted to see what an AI tool would look like if it were designed not to do the thinking for you but to expand your thinking.

Prompted generates only questions: oblique provocations designed to reframe, complicate, or redirect thinking rather than resolve it. Users bring something they’re sitting with–a piece of writing, an image, a question–and place it on an open canvas. To initiate a consultation, they hold rather than click. Where most AI interfaces make it easy to fire off a prompt without much thought because you can always revise your input or regenerate the output, the hold here asks you to mean what you place. What follows is a delay you cannot predict or plan around. When the system is ready, a second hold is required to receive the response: three to five oblique questions that appear as connected nodes on the canvas. These questions cannot be regenerated or revised. What appears is what you work with. To get anything further, you have to bring more of your own thinking.

The canvas grows as a spatial record of this process. Unlike chat interfaces that scroll past earlier exchanges, the canvas preserves everything, including branches, dead ends, returns. Over time it becomes a map of where your thinking has been: not a log of answers received but a topology of questions encountered. The thinking stays with the user. The system provides provocations; you do the sensemaking.

What AI adds
The I Ching’s hexagrams are fixed. The DEW Line deck and Oblique Strategies draw from the same cards regardless of what you’re working through. Their generality is part of what forces the interpretive work onto you. But AI adds something these systems can’t: responsiveness to the specific language you bring.

Large language models have no genuine understanding, but they are good at pattern-matching and rearranging your language into something not quite what you meant, angled in a direction you didn’t intend. For most AI applications, that’s a problem to engineer around; you want the model to be correct, relevant, reliable. For sensemaking, those aren’t necessarily the point. The provocation doesn’t need to be “correct.” It doesn’t even need to be relevant in any obvious way. It just needs to be oblique enough to make you think again about what you meant, and AI is adept at creating a kind of strangeness informed by the language you offer it.

The lineage from the I Ching through Oblique Strategies to tools like Prompted suggests that an AI interface can be designed for something other than speed or correctness; it can be designed to keep a question open, to give you material that sustains your thinking rather than replaces it. The kind of thinking that needs open time, that builds through sitting with uncertainty, that arrives somewhere you couldn’t have planned for.

Meta's reported efforts to develop facial-recognition capabilities for future smart glasses have renewed concerns about privacy, biometric surveillance, and anonymity in public space. According to multiple reports, the company has explored technology that would allow AI-powered glasses to identify individuals viewed through a built-in camera and provide information about them to the wearer. While Meta's current Ray-Ban Meta smart glasses do not include facial recognition, the possibility of adding the technology has prompted criticism from privacy advocates and civil liberties organizations.

The concern is not simply that smart glasses can capture images. Smartphones already do that. Facial recognition introduces a new layer of automated identification, allowing cameras to recognize individuals in real time. Organizations including the American Civil Liberties Union (ACLU) and the Electronic Privacy Information Center (EPIC) have warned that wearable facial recognition could enable individuals to be identified without their knowledge or consent. Critics argue that such systems risk normalizing biometric surveillance in everyday environments, from public transit and restaurants to classrooms, workplaces, and cultural institutions.

Unlike fixed security cameras, smart glasses move with their users and operate within ordinary social interactions. The technology raises questions about whether people should retain a reasonable expectation of anonymity when moving through public spaces. It also highlights broader concerns regarding how biometric data is collected, stored, and protected.

The debate arrives as technology companies increasingly position smart glasses as a next-generation computing platform powered by artificial intelligence. Current devices can already identify objects, translate text, and answer questions about the wearer's surroundings. Facial recognition would extend that capability to human identity itself. The issue is not whether facial-recognition technology is possible. The debate centers on whether biometric identification should become a routine feature of consumer wearable devices. As smart glasses continue to evolve as AI platforms, regulators, technology companies, and civil society groups are increasingly confronting that question.

Dubai has announced plans for the Museum of Digital Art (MODA), a new institution dedicated to digital culture, immersive media, and emerging technologies. The project was officially launched by Sheikha Latifa bint Mohammed bin Rashid Al Maktoum in May 2026 and will be developed within the DIFC Zabeel District as part of Dubai International Financial Centre’s long-term cultural strategy.

According to the official announcement, MODA will span five floors and include permanent and temporary exhibitions, immersive installations, educational programming, and research initiatives focused on digital and computational art practices. The museum also plans to incorporate a “digital twin” platform intended to extend parts of the institution into virtual environments accessible remotely.

The building is being designed by Adrian Smith + Gordon Gill Architecture, the Chicago-based firm behind projects including the Burj Khalifa and Jeddah Tower. Dubai Culture and Arts Authority will oversee the museum’s cultural direction alongside DIFC.

While the announcement positions MODA as the region’s first dedicated museum for digital art, several operational details remain undisclosed, including the opening date, inaugural exhibitions, curatorial leadership, and acquisition strategy. At present, the project exists primarily as an institutional and architectural proposal rather than a fully articulated museum program.

The launch arrives amid broader regional investment in technology-centered cultural infrastructure. Abu Dhabi recently opened teamLab Phenomena Abu Dhabi, while Saudi Arabia’s Diriyah Art Futures has expanded programming around digital and new media practices. MODA adds Dubai to a growing network of Gulf institutions positioning computational culture, immersive systems, and digital production as central components of future cultural development.

The museum’s long-term significance will likely depend less on immersive spectacle than on whether it develops a sustained curatorial and research framework capable of supporting software-based, networked, and AI-driven artistic practices over time.

South Korean artist Beak Jungki works across sculpture, photography, material research, and environmental systems. In his long-running is of series, Jungki extracts pigments from leaves, flowers, and plant matter collected directly from the landscapes he photographs, producing images that remain materially connected to the environments they depict.

First initiated in 2007, the is of series includes works such as is of Naejangsan, is of Palgongsan, is of Dumulmeori, is of Boseong Green Tea Field, and is of Songnisan. Because the organic pigments gradually fade and oxidize over time, the works treat photography as a changing material process rather than a fixed visual archive. Acrylic chambers, oxygen-removal devices, and environmental control systems slow that transformation without fully stopping it, connecting the work to broader conversations around embodiment, biosensing, and post-digital creative practice.

Printing Landscapes With Their Own Material
Jungki’s process begins with site-specific collection. In mountain regions such as Naejangsan, Palgongsan, and Songnisan National Parks in South Korea, he gathers fallen leaves and other organic material directly from the landscapes he photographs. These materials are processed into pigments and used in inkjet printing systems to produce the final image.

The process changes the photograph from a stable recording into a material extension of the landscape itself. Conventional photography translates reflected light into industrially standardized pigments or pixels. Jungki instead constructs images from material extracted directly from the sites being depicted.

Works such as is of Naejangsan 2024-5 and is of Palgongsan 2024-3 carry visible traces of seasonal variation through pigments extracted from autumn foliage. The colors are chemically derived from the landscape itself, introducing instability into the image from the moment it is produced. Organic pigments gradually shift through exposure to oxygen, humidity, temperature, and light.

One of the clearest examples is is of Naejangsan 2023-1, which was intentionally exhibited without a preservation chamber. By the time the work reappeared in later exhibitions, substantial fading had already occurred. The altered surface became evidence of elapsed time and material transformation rather than damage requiring restoration.

Preservation as Artistic Infrastructure
Many of the is of works are displayed within acrylic chambers containing oxygen-removal devices, nitrogen systems, vacuum pumps, and environmental controls. These mechanisms are not hidden. Tubes, pumps, hoses, and sealed structures remain visibly integrated into the installations.

In works such as is of Songnisan 2024-4, the preservation apparatus becomes inseparable from the image itself. Acrylic chambers and environmental control systems regulate oxidation and slow the fading of the organic pigments, making the conditions required to sustain the image part of the work itself. The installation combines photographic display, preservation technology, and environmental control into a single system.

The systems operate quietly in the background, managing atmosphere, exposure, and decay. This shifts attention away from technological novelty and toward the material conditions shaping how the image changes over time. The chamber systems frame preservation as an ongoing environmental process rather than a permanent solution. The pigments continue to fade, even under controlled conditions, keeping the focus on duration, transformation, and the instability of the image itself.

Biological Time and the Post-Digital Image
The is of series ultimately challenges one of photography’s core assumptions: permanence. Since the nineteenth century, photographic technologies have largely pursued stabilization, reproducibility, and archival durability. Jungki moves in the opposite direction. His images are designed to change.

Jungki’s installations do not present nature as scenery to be consumed visually. Instead, they stage direct encounters with material change over time. The images slowly fade while pumps, sealed chambers, and oxygen-removal systems work continuously to delay their deterioration. The chambers slow entropy but cannot eliminate it. The image remains temporary.

Instability becomes part of the image itself. Atmosphere, chemistry, maintenance, and biological decay all shape how the works continue to evolve after production. The photographs are therefore less about capturing a landscape than participating in its ongoing material life. Mountains become pigments. Pigments become images. Images gradually return to environmental process. By embedding ecological transformation directly into photographic production, Beak Jungki expands what a photograph can be: not a fixed representation of the world, but a material system continuously shaped by biological time.

Ava Hudson’s work moves between sculpture, wearable systems, biosensing, and embodied interaction. Using textiles, circuitry, hardware materials, and sculptural forms, she builds objects that examine how bodies move through physiological, technological, and environmental systems. In works that use electronics, such as Pulse, sensors and circuitry remain visibly embedded within the wearable structure rather than disappearing behind seamless fabrication. Drawing from backgrounds in cognitive science and architecture, Hudson approaches wearable technology as sculptural and spatial practice. This takes different forms across works including Pulse, a heartbeat-responsive wearable that transforms cardiac activity into light; Lucent Body, an ornamented wearable sculpture layered with materials associated with adornment, industry, and nature; and Anti-Surveillance Helmet, a speculative wearable examining visibility, monitoring, and bodily exposure within networked environments.

The Heartbeat as Raw Material
At the center of Pulse is a relatively simple interaction. A heartbeat sensor connected to a custom circuit board causes light embedded within the wearable object to fluctuate in real time according to changes in the wearer’s pulse. The project is built from LEDs, magnetic sensors, recycled circuitry, crystals, textiles, and 3D-printed brain hemispheres.

Circuit boards and electronic components remain exposed rather than concealed beneath polished fabrication. Scrap computer chips sit directly against fabric and sculptural surfaces. Sensors are embedded into the wearable structure itself rather than separated into hidden technical layers. The result feels closer to a constructed artifact than a finished consumer device. Nothing disappears into seamlessness. Light moves unevenly across the object’s surface as the body shifts and the pulse changes.

There are no dashboards, biometric readouts, or quantified visualizations translating the body into data streams. The heartbeat appears instead as changing luminosity moving through material form. That distinction gives the work much of its tension. In Hudson’s practice, biosignals behave less like information than raw material.

Wearables After the Wellness Interface
Most contemporary biosensing technologies are built around optimization. Smart watches measure recovery scores, sleep cycles, heart rate variability, and physical performance, converting bodily activity into measurable outputs designed for monitoring and improvement.

Hudson’s wearables do not frame the body as something to quantify or refine. Instead, sensing systems trigger changes in light, texture, and spatial response that remain perceptual rather than informational. That relationship between body and material also appears in Lucent Body. The wearable sculpture expands outward from the body through layered forms and materials associated with adornment, industry, and nature. Its armor-like structure frames the body through protection, visibility, and physical presence rather than technological seamlessness.

Fabric plays an important role throughout this work. Textiles soften the hard visual logic typically associated with electronics and sensing systems, introducing tactility and bodily proximity into structures built from circuitry and light. The materials never fully resolve into either garment or machine.

The same tension runs through Anti-Surveillance Helmet, a speculative wearable examining the aesthetics and politics of visibility within networked surveillance systems. The helmet evokes both protective equipment and technological apparatus, exaggerating the body’s relationship to monitoring and exposure rather than attempting to hide it.

Like Pulse, the work keeps its technological structure materially legible. Electronic components remain visible within the object instead of disappearing behind frictionless industrial design. Hudson’s wearables consistently resist the visual language of seamless consumer technology, opting instead for layered surfaces, exposed systems, and visible construction.

Embodiment, Exposure, and Responsive Systems
Hudson’s work connects to longer histories of wearable sculpture, embodied art practices, and technologically informed material experimentation. Sensors, circuitry, textiles, and bodily movement operate together within the same physical structure. Hudson does not stage technology as spectacle. The interactions remain relatively minimal: pulses of light, exposed circuitry, responsive surfaces, shifting material conditions. Attention stays focused on the relationship between the body and the object rather than on immersive technical effects.

Across these projects, the body operates as both subject and site — shaped by systems of sensing, visibility, protection, and technological mediation. Physiological activity becomes visible through light and material response without collapsing entirely into metrics or data visualization. Her work does not reject sensing technologies, nor does it romanticize them. Instead, Hudson treats responsive systems as material structures capable of shaping perception through light, circuitry, touch, and bodily presence.
 
Hudson’s upcoming solo exhibition will open June 8 at Gallatin Galleries in New York City. Centered on wearable art and technology, the exhibition explores the body as a conduit between human, technological, and ecological systems — a theme running consistently throughout Hudson’s practice.

Presented by Fellowship and ARTXCODE at Art Basel Hong Kong 2026, RECURSIONS 遞迴 marks a shift in Sougwen Chung’s practice. Known for live, co-performative drawing systems, Chung extends this inquiry into painting, robotics, and light-based installation, positioning gesture as the result of feedback between human and machine.

At the center of the exhibition is RECURSION 0 (2026), a monumental scroll produced through a kinetic robotic system trained on the artist’s accumulated drawing data. Biosensor input modulates the system in real time, allowing machine movement to extend and echo human gesture across linen. The surface records this exchange as intention and response accumulate without a clear point of origin.

Six related works, RECURSION 1–6 (2026), translate this durational process into discrete paintings. Rather than isolating authorship, the works compress it. Marks layer without hierarchy, holding both the initiating gesture and its machinic continuation within a single field.

The investigation expands in RECURSION Dataset 1314 (For a Lifetime) (2026). Drawing and brainwave data are transposed into a suspended LED mesh, where patterns evolve continuously. The title references the Chinese homophone 一三一四, echoing 一生一世 (“one lifetime”), framing attention as duration—extended, compressed, and rendered as light.

Across these works, Chung develops a framework for co-creation across human, machine, and environment, where authorship is distributed through sensing, training data, and system response.

This exhibition builds on Chung’s Drawing Operations series (initiated in 2015), which has traced the evolution of human–machine collaboration through robotic drawing, neural networks, and biofeedback systems. In RECURSIONS, that trajectory consolidates into works where gesture is continuously negotiated across surfaces, systems, and time.

OpenAI has confirmed it will discontinue Sora, its text-to-video app, along with the associated API used by developers and studio partners. The decision arrives within a year of Sora’s rollout as a consumer-facing product and reflects a broader consolidation of OpenAI’s product strategy ahead of a potential IPO.

The shutdown points to a structural issue: high-fidelity generative video has advanced rapidly at the model level but remains difficult to sustain as a standalone product. Third-party data from App figures shows Sora downloads declining from 3.3 million in November 2025 to 1.1 million by February 2026, suggesting limited retention beyond initial experimentation.

The gap between technical capability and practical use is not new. Projects such as Sunspring by Ross Goodwin demonstrated early interest in machine-generated cinema, while recent systems from Runway and Pika Labs have made video synthesis widely accessible. Yet integration into production workflows—editing, continuity, rights management—remains fragmented.
OpenAI’s internal reallocation reinforces this shift. According to reporting, Sora’s research team will move toward “world simulation” work linked to robotics, suggesting that video generation may increasingly function within broader simulation systems rather than as a standalone consumer tool.

The move also coincides with stronger-performing segments. OpenAI’s Codex has surpassed $1 billion in annualized revenue, underscoring where immediate utility and monetization are clearer.
Sora’s discontinuation does not mark the end of generative video. Instead, it highlights its current position: technically advanced, but still searching for a stable role within everyday creative and industrial pipelines.

The 2026 Whitney Biennial, curated by Marcela Guerrero and Drew Sawyer, has been framed by critics as a survey shaped by derivative tendencies—what one recent New Yorker review described as “ChatGPT art,” where works appear as “facsimiles of facsimiles” detached from clear artistic lineage. The critique targets a perceived lack of acknowledgment of influence, but the broader issue extends beyond authorship. It points to a shift in how cultural production operates under conditions increasingly aligned with machine learning.

Appropriation has long been central to contemporary art. Sherrie Levine’s After Walker Evans (1981) and Elaine Sturtevant’s reworkings of canonical artists foregrounded copying as a critical act, making lineage explicit and conceptually necessary. By contrast, many works in the current Biennial have been read as operating through resemblance without articulation. One critic has pointed to visual and conceptual echoes between artists such as Mo Costello and earlier photographic practices associated with James Welling, or between Nour Mobarak’s installations and the environmental systems of Pierre Huyghe.

What distinguishes this moment is not the presence of influence, but its flattening. Machine learning systems are trained on large-scale datasets where attribution is obscured and context is compressed. That logic appears to be mirrored in some of the works discussed by critics: influence circulates as aesthetic pattern rather than as a situated, critical relationship. The result is work that can resemble prior practices without engaging their conditions of production.

Not all artists in the exhibition operate this way. Agosto Machado’s assemblages draw directly on the history of downtown New York performance and activism, functioning as material archives rather than stylistic references. Similarly, Sula Bermúdez-Silverman’s glass and steel sculptures engage Dada and postwar sculpture through explicit formal and historical dialogue.

The Biennial does not present a unified position on contemporary art, but it does register a shift. As generative systems reshape how images, forms, and styles circulate, the question is less about whether artists use AI and more about whether artistic practice can maintain legible relationships to its sources. In this context, authorship becomes less about originality than about the capacity to situate work within a field of references that remains visible rather than flattened into output.

When the New Museum reopened its expanded building on March 21, 2026, it did so with New Humans: Memories of the Future, a large-scale exhibition spanning the museum and examining how conceptions of the human have shifted across the twentieth and twenty-first centuries. The exhibition brings together more than two hundred artists, writers, scientists, architects, and filmmakers, placing contemporary works in dialogue with historical material.

Curated by Massimiliano Gioni, Gary Carrion-Murayari, Vivian Crockett, and Madeline Weisburg, the exhibition does not limit the idea of “new humans” to recent developments in artificial intelligence or biotechnology. Instead, it situates current technological conditions within a longer history of artistic and scientific attempts to represent the body, perception, and human identity. Materials from early twentieth-century avant-garde movements appear alongside contemporary works, reflecting how earlier moments of industrialization, mechanization, and media transformation produced new ways of imagining the human. By placing these periods in relation, the exhibition frames present-day concerns—including machine learning and computational media—within a broader continuum of technological and cultural change.

A Century-Long View of the Human Under Pressure
One of the exhibition’s defining structural moves is its refusal of a linear timeline. The show “traces a diagonal history” of the twentieth and twenty-first centuries, placing canonical figures such as Constantin Brâncuși, Salvador Dalí, Man Ray, and August Sander in proximity to artists including Bruce Lacey, Rammellzee, Toyen, and Unica Zürn. This juxtaposition situates modern and contemporary works within a shared field rather than a chronological sequence.

Scientific imagery is integrated directly into this structure including Lennart Nilsson’s embryo photographs, Santiago Ramón y Cajal’s drawings of neural structures, Franz Tschackert’s Glass Man, and Wilder Penfield’s sensory homunculus. Presented alongside artworks, these materials reflect how scientific models and visual culture have both contributed to representations of the human body. The Guardian described the exhibition as a 732-object survey of art, artifacts, and visual culture, noting the prominence of historical material such as Nilsson’s fetal photography from the 1960s, produced using endoscopic imaging techniques.

Contemporary Commissions and Expanded Bodies
The exhibition includes more than fifteen new commissions by contemporary artists including Ryan Gander, Camille Henrot, Jamian Juliano-Villani, Wangechi Mutu, Hito Steyerl, Alice Wang, and Santiago Yahuarcani. These works are presented alongside historical material as part of the museum’s reopening program.

Several individual works clarify the exhibition’s range of approaches to the body and its representation. For example, Aneta Grzeszykowska’s photographic series Mama (2018), in which the artist uses a lifelike silicone model of her own body to construct staged images of familial relationships. The exhibition also includes Goshka Macuga’s projects Before the Beginning and After the End (2016) and Transhumanism (2017), which combine archival material and machine-assisted drawing. Nearby is Alina Szapocznikow’s Kaprys – Monstre (1967), composed of cast body fragments in synthetic materials.

Other sections extend beyond the individual body to speculative environments. The curatorial framework references imagined cities and habitats by Sophia Al-Maria, Bodys Isek Kingelez, Gyula Kosice, Constant Nieuwenhuys, Hariton Pushwagner, and Albert Robida. It also includes Anicka Yi’s floating aerobes, originally developed for Tate Modern’s Turbine Hall, presented here within the New Museum galleries.

Positioning the Present Within a Longer Continuum
The exhibition is positioned within ongoing discussions about artificial intelligence and technological change. Rather than focusing exclusively on recent developments, New Humans: Memories of the Future situates these topics within a longer history of visual culture, scientific representation, and artistic experimentation.

As the inaugural exhibition for the New Museum’s expansion, the show operates at the scale of a survey, combining historical material with contemporary commissions. This structure places current technological conditions in relation to earlier moments of industrialization, mechanization, and media transformation.

Across its sections, the exhibition assembles works that address recurring questions about representation, identity, and the relationship between humans and technical systems. By bringing together materials from different disciplines and time periods, it frames the future as something shaped through existing cultural, scientific, and artistic frameworks rather than as a distinct break from the past.

The process typically begins with physical collage. Fragments—often drawn from photographic sources, printed graphics, or hand-produced marks—are cut, arranged, and reassembled into composite images. These materials are then digitized, allowing Hale to manipulate scale, layering, and movement within animation software.

Rather than treating collage as a static medium, Hale uses animation to destabilize the image. Elements repeat, loop, and shift position, producing compositions that remain in flux. Figures fragment and recombine, backgrounds fold into foregrounds, and visual hierarchies are continuously restructured. This approach places Hale’s work within a broader shift in image-making, where the boundary between physical and digital processes is increasingly fluid. The collage is not a finished object but an intermediate state within a longer computational workflow. Scanning and editing extend the material logic of cutting and layering into time-based media.

Formally, the work often relies on density and accumulation. Multiple visual registers—texture, pattern, figure, and color—coexist within the same frame. Repetition plays a central role, both as a compositional strategy and as a temporal device in animation. Loops and cycles create rhythm, while small variations prevent the image from resolving into a stable configuration.

Hale’s animations do not follow linear narrative structures. Instead, they operate through transformation and association. Meaning emerges through juxtaposition, scale shifts, and the reconfiguration of familiar visual elements into unfamiliar arrangements.

In this sense, the work can be understood as a system of image operations: cutting, layering, duplicating, and animating. Each stage of the process introduces new relationships between elements, allowing the composition to evolve across both space and time. By combining tactile, material processes with digital animation, Hale’s practice foregrounds the constructed nature of images. The viewer encounters not a seamless visual surface, but a composition that reveals its own assembly—where fragments remain visible even as they move.

Founded in 2009 by Andrew Zolty, BREAKFAST operates at the intersection of industrial design, software engineering, and kinetic art. Based in New York, the studio develops large-scale installations that translate digital data and human interaction into coordinated physical motion. These projects function as engineered systems composed of actuated elements rather than screen-based displays or static objects.

Instead of rendering information through light-based pixels, BREAKFAST constructs environments where patterns and behaviors emerge through mechanical movement. Its work contributes to a growing body of practice exploring how computation can be embedded directly into physical systems.

Mechanical Pixels and Data-Driven Motion
At the core of BREAKFAST’s practice is a studio-developed approach to modular kinetic units—often described as “mechanical pixels.” These components function similarly to digital pixels but operate through motors and custom mechanical linkages rather than light emission.

Projects such as Brixel Mirror use mirrored elements that rotate or tilt in response to input. When arranged in large grids, these units generate shifting fields of reflection, turning architectural surfaces into responsive systems. Visual resolution is determined by the number and precision of actuated elements rather than screen density. The studio has also adapted flip-disc technologies—historically used in transit signage—into large-scale installations. These systems transition between states through coordinated mechanical motion, producing patterns and imagery through synchronized physical change.

Many installations are connected to live or continuously updating data sources. BREAKFAST has developed works that respond to financial data, including cryptocurrency markets, as well as environmental inputs such as weather. These data streams are translated into motion across kinetic surfaces, allowing information to be perceived through changes in rhythm, density, and movement rather than through conventional charts or displays. This approach shifts data from representation to behavior. Instead of presenting information for direct interpretation, the systems operate continuously, embedding data into the surrounding environment.

Architectural Integration and Continuous Operation
BREAKFAST’s installations are typically commissioned for corporate headquarters, cultural institutions, and other public-facing environments. Rather than functioning as standalone objects, they are integrated into walls, façades, or interior surfaces, operating as dynamic architectural elements.

Scale is a defining characteristic. Many projects occupy large surfaces, where coordinated movement alters the perception of space over time. The experience extends beyond visual output: the sound of motors, changing reflectivity, and temporal variation all contribute to how the work is perceived. Because these systems are deployed in high-traffic environments, durability and maintainability are central to their design. Installations are built for continuous operation, requiring robust fabrication, modular components for repair, and software systems capable of long-term stability.

Some projects incorporate sensors that respond to viewer presence or interaction, enabling real-time adjustments. Others operate autonomously, driven by external data streams. In both cases, the installations function as responsive systems embedded within their environment rather than static displays.

A Hybrid Studio Model and the Future of Physical Interfaces
BREAKFAST’s organizational structure reflects the complexity of its work. The studio integrates mechanical engineering, electronics, software development, industrial design, and fabrication within a single workflow. Projects are developed end-to-end, from concept and prototyping through manufacturing and installation, allowing technical and aesthetic decisions to be addressed simultaneously.

Under Zolty’s leadership, the studio operates across art, design, and commercial production. Many projects are commissioned at architectural scale, while maintaining a consistent focus on how physical systems can embody computation. This positioning has placed BREAKFAST within ongoing discussions around kinetic art, responsive environments, and spatial interfaces. Rather than replacing conventional screens, the studio extends computation into physical space. Digital inputs are translated into coordinated movement across surfaces, producing systems that operate continuously and often without direct interaction.

In this model, information is not confined to displays but distributed across material systems. Movement becomes a primary means of output, and environments take on a form of behavioral responsiveness shaped by the interaction of hardware, software, and data.

For much of the past century, air conditioning, automated shading, and sensor-driven building management systems have controlled temperature, airflow, and light across much of the built environment. Yet architecture has long incorporated another approach: designing buildings so that environmental regulation emerges from their physical structure. Through orientation, airflow, material behavior, and solar shading, buildings can moderate interior conditions without relying entirely on mechanical equipment.

Climate-responsive design has deep historical roots. Long before mechanical heating and cooling systems became widespread in the twentieth century, buildings responded directly to local environmental conditions. Courtyard houses created shaded microclimates, thick masonry walls stabilized temperature swings, and ventilation towers encouraged air movement through interior spaces. Today architects often combine these established strategies with environmental modeling tools that simulate airflow, solar exposure, and thermal performance during the design process.

Designing with Air: Passive Ventilation Strategies
Air movement remains one of the most effective ways to moderate indoor temperature. Long before mechanical ventilation systems existed, architects developed ways to guide airflow through buildings using orientation, spatial layout, and carefully positioned openings.

Cross ventilation is among the most widely used strategies. Openings placed on opposing sides of a building allow prevailing winds to move through interior spaces, carrying away heat and introducing cooler outdoor air. When combined with shading and appropriate building orientation, cross ventilation can substantially reduce indoor heat accumulation in warm climates.

Buildings can also generate vertical airflow through the stack effect. Because warm air rises, structures can incorporate atriums, chimneys, or ventilation shafts that allow heated air to escape at higher elevations. As warm air exits, cooler air is drawn in through lower openings, producing continuous circulation driven by temperature differences rather than mechanical fans.

Contemporary architects continue to refine these principles. Architect Francis Kéré’s Gando Primary School in Burkina Faso, first completed in 2001, uses compressed earth walls and a raised metal roof structure separated from the ceiling below. The gap between roof and ceiling allows heat to accumulate and escape while shaded openings draw cooler air through the classrooms. In regions where wind conditions vary, buildings may incorporate ventilation towers or high-level openings designed to capture breezes and direct them into interior spaces. These systems operate through pressure differences created by wind and temperature gradients rather than electrical equipment.

Thermal Mass and the Physics of Materials
Another key strategy in climate-responsive architecture involves the thermal behavior of building materials. Materials such as stone, adobe, rammed earth, and concrete have high thermal mass, allowing them to absorb heat slowly and release it gradually over time.

In hot climates, this property moderates indoor temperatures by delaying the transfer of external heat into interior spaces. Thick earthen or masonry walls absorb heat during the day and release it later in the evening as outdoor temperatures fall, reducing temperature fluctuations between day and night. Architects continue to explore these material properties in contemporary construction. Anna Heringer’s METI Handmade School in Rudrapur, Bangladesh, completed in 2006, combines earthen walls with bamboo structural elements derived from regional building traditions. The thick earthen walls provide thermal mass that slows heat transfer, while permeable bamboo upper levels and elevated openings allow air to circulate through the building.

Thermal mass often works in combination with ventilation strategies. When buildings cool at night through natural airflow, heavy materials release heat stored during the day. This daily cycle of heat absorption and release allows the structure itself to participate in regulating interior climate.

Controlling Solar Radiation Through Architecture
Reducing solar heat gain before it enters a building is often more effective than removing it after interior temperatures rise. Climate-responsive architecture therefore addresses solar radiation directly through façade design, shading systems, and spatial depth.

Shading devices such as overhangs, screens, and vertical fins limit direct sunlight on walls and windows, reducing the amount of heat absorbed by the building envelope. In many climates, carefully designed shading strategies can significantly reduce cooling demand while still allowing daylight into interior spaces. The Institut du Monde Arabe in Paris, designed by Jean Nouvel and Architecture-Studio and completed in 1987, remains one of the most widely cited architectural experiments in solar control. Its south-facing façade incorporates hundreds of mechanical diaphragms inspired by traditional mashrabiya screens. Photoelectric sensors adjust the apertures in response to changing light conditions, regulating daylight and solar gain.

More recent buildings often rely on simpler passive systems. Deep window recesses, perforated screens, and layered façades create shaded buffer zones that limit solar exposure while allowing filtered daylight to reach interior spaces.
Architect Marina Tabassum’s Bait Ur Rouf Mosque in Dhaka, completed in 2012, provides a contemporary example. Constructed primarily from brick, the building uses thick masonry walls, carefully positioned openings, and natural ventilation to moderate interior conditions. Light enters through perforated brick surfaces and roof apertures, producing diffuse illumination while limiting direct solar heat gain.

Reintegrating Environmental Intelligence into Building Design
The renewed interest in climate-responsive architecture reflects broader shifts in how buildings are designed and operated. According to the United Nations Environment Programme’s Global Status Report for Buildings and Construction, buildings and construction together account for roughly 37 percent of global energy-related carbon dioxide emissions, much of it associated with heating and cooling.

Digital tools remain central to the design process. Architects routinely use environmental simulation platforms such as EnergyPlus and Ladybug Tools to model airflow, solar exposure, and thermal performance. The resulting buildings, however, regulate environmental conditions through physical relationships—air movement, thermal mass, and solar shading—rather than electronic control systems. This approach also draws on architectural traditions developed long before modern mechanical infrastructure. Courtyard houses across the Middle East and North Africa create shaded microclimates and encourage airflow. Windcatchers in Iran and the Gulf channel prevailing breezes into interior spaces. Earthen construction in arid regions uses thermal mass to moderate daily temperature swings.

Contemporary architects are not simply replicating historical forms. Instead, they reinterpret these environmental principles using modern materials, structural systems, and computational analysis. Passive design strategies are increasingly integrated with contemporary construction techniques and performance modeling. As climate pressures intensify and energy demand rises, buildings that regulate temperature, airflow, and light through their own physical properties are attracting renewed attention. In these projects, environmental control is not a technological layer added after construction. It is embedded directly in the design logic of the building itself.

Generative AI images are typically produced through prompts—the short text instructions used to guide models such as Stable Diffusion. But many artists working with machine learning are operating at a different level of the technology. Rather than focusing on prompts alone, they design the datasets, training processes, and computational pipelines that shape how generative systems produce images. The release of open diffusion models accelerated this shift. Stable Diffusion, introduced in 2022 by Stability AI and collaborators, made it possible to run high-quality generative image models locally rather than through closed platforms. Because the model weights are publicly available, artists can download, modify, and retrain the system using their own image datasets. They can then integrate those models into custom software environments and generation workflows.

As a result, generative AI artists increasingly resemble software developers and research engineers. Studio practice now often includes dataset construction, model training, parameter experimentation, and code development. In this context, the artwork is not simply the image produced by the model—it is the system that generates it.

Building Diffusion Workflows Instead of Single Images
Active in the machine learning art community since the late 2010s, Helena Sarin approaches neural networks as creative instruments that can be trained, modified, and integrated into custom production workflows. Her practice often involves assembling curated image datasets, training generative models, and experimenting with inference pipelines that shape how images are produced. Earlier projects used generative adversarial networks (GANs) trained on everyday visual subjects such as food, household objects, and hand-drawn sketches. By controlling the structure of the dataset and the training process, Sarin investigates how neural networks transform familiar visual categories into stylized or abstract representations.

Rather than relying solely on text prompts, Sarin’s workflow frequently involves modifying models and experimenting with different generation parameters to explore how neural networks interpret visual information. This approach treats the generative model not simply as a tool for producing images but as a system whose behavior can be shaped through training data and software configuration. Sarin regularly shares experiments and technical insights through open platforms such as GitHub and social media, contributing to a broader ecosystem of artists working directly with machine learning systems. The practice reflects a growing tendency among generative AI artists to design the computational pipelines behind image generation rather than focusing only on the outputs.

German artist Mario Klingemann represents an earlier phase of this system-oriented approach to machine learning art. Beginning in the mid-2010s, Klingemann trained generative adversarial networks (GANs) on large image datasets, experimenting with how neural networks reinterpret visual culture. Klingemann’s experiments demonstrated how artists could treat neural networks themselves as creative materials. Rather than simply generating images with existing tools, he trained models and explored how generative systems transform visual patterns contained within large image collections. Although GAN-based systems differ technically from diffusion models, the artistic strategy remains closely related. The artist constructs a training environment, defines the dataset, and observes how the model learns visual structures from those inputs.

Projects such as Memories of Passersby I (2018), a machine learning installation that continuously generates synthetic portraits, demonstrate how generative systems can function as autonomous image-making processes. In this approach, the behavior of the system itself becomes a central artistic outcome.

Dataset Construction as Artistic Practice
The most consequential artistic decision occurs long before image generation begins. It happens during dataset construction. Argentine artist Sofia Crespo, working both independently and through the collective Entangled Others, has developed a machine learning art practice centered on biological imagery. Her work explores how neural networks interpret patterns found in natural forms such as coral structures, microscopic organisms, and plant morphology. Rather than relying exclusively on pre-trained models, Crespo often assembles custom datasets derived from biological imagery and scientific references. These datasets guide how machine learning models learn visual structures.

In several projects, Crespo explores how generative systems can synthesize organisms that appear biologically plausible while remaining entirely artificial. The results often resemble unfamiliar species—creatures that seem drawn from evolutionary processes but exist only within the representational logic of neural networks. In this approach, the dataset functions as a form of artistic material. The images used for training determine what the model can learn, and therefore what kinds of forms it can generate. Dataset curation becomes a central component of the artistic process. This method has become increasingly common in generative AI art, where artists treat training data not as neutral input but as a medium shaped through selection, filtering, and categorization.

Neural Networks as Research Instruments
British artist and researcher Terence Broad approaches generative AI through a slightly different lens. His work often treats neural networks as tools for investigating how machine learning systems interpret visual culture. Broad completed a doctoral project at Goldsmiths, University of London that examined machine learning models as analytical instruments within media practice. One of his widely discussed works, Autoencoding Blade Runner, involved training an autoencoder neural network on the entire 1982 film Blade Runner. After learning a compressed representation of the film, the system attempted to reconstruct each frame.

The resulting video reveals how neural networks abstract visual information. Faces blur into geometric shapes, environments dissolve into color fields, and cinematic details collapse into simplified forms. The project functions both as an artwork and as a study of how machine learning systems perceive images. Broad’s later work has engaged with generative models and related machine learning techniques, examining the ways these systems transform visual media during training and inference. By building experimental training pipelines and observing their outputs, he investigates the assumptions embedded within computational vision systems. This approach highlights another dimension of machine learning art practice. Generative models are not only image-making tools. They can also operate as research instruments that reveal how artificial intelligence systems encode and reconstruct visual information.

Across these practices, a broader shift becomes visible. Generative AI artists are no longer defined primarily by how they prompt models. Instead, they design the environments in which models operate. Datasets, training pipelines, and inference workflows form the core of the studio. Images become traces of larger computational systems. As open-source tools continue to expand access to diffusion models and machine learning frameworks, the role of artists as system builders is likely to grow. In this emerging landscape, creative practice increasingly involves designing the mechanisms that generate images rather than producing those images directly.

For most of the industrial era, movement in machines has depended on rigid structures and mechanical assemblies. Motors rotate shafts. Hinges define joints. Control systems coordinate the interaction of components. Soft robotics proposes a different model: movement emerging directly from the physical behavior of materials. Instead of rigid frames, soft robotic systems rely on elastomers, flexible polymers, hydrogels, and composite materials capable of bending, swelling, contracting, or deforming in controlled ways.

Geometry, material composition, and fabrication techniques encode behavior into the structure itself. In these systems, movement is not produced by mechanisms layered onto materials; it arises from the materials themselves. Soft robotics research sits at the intersection of robotics, materials science, and advanced fabrication. Researchers increasingly treat matter not as a passive substrate but as an active element capable of performing sensing, actuation, and limited forms of control.

Magnetic Matter and Remote Actuation
One of the most active areas of soft robotics research explores magnetic control of flexible materials. Scientists embed microscopic magnetic particles into elastomers or polymer matrices, allowing external magnetic fields to manipulate the resulting structures without onboard motors or rigid mechanisms. Roboticist Metin Sitti—formerly director of the Physical Intelligence Department at the Max Planck Institute for Intelligent Systems and now based at Koç University—has led influential research into magnetically actuated soft robots. These devices are typically fabricated from silicone or similar elastomer materials mixed with magnetic particles. When exposed to controlled magnetic fields, the robots can bend, crawl, roll, or swim through complex environments.

Research from Sitti’s group demonstrates millimeter-scale soft robots capable of navigating confined spaces and uneven terrain through externally applied magnetic fields. Because the robots do not require onboard motors or batteries, they can be fabricated at extremely small scales. The approach has attracted interest for potential medical applications such as targeted drug delivery or minimally invasive procedures, where small flexible devices must move through delicate environments. The significance lies less in the specific devices than in the design logic they represent. Actuation becomes a property of the material system itself. Magnetic particles distributed throughout a flexible structure allow movement to emerge through interaction with external fields rather than mechanical joints.

Synthetic Muscle and Liquid Crystal Elastomers
Another major direction in soft robotics research and materials science involves materials capable of contracting in response to energy input. Liquid crystal elastomers (LCEs) combine the molecular alignment properties of liquid crystals with the elasticity of polymer networks. When heated or illuminated, these materials contract along a programmed axis, producing motion analogous to biological muscle. Researchers including materials scientist Timothy J. White at the University of Colorado Boulder have developed LCE actuators that respond to light or heat. In these systems, thin polymer films are fabricated with aligned molecular structures. When exposed to stimuli such as infrared light, the material contracts and bends in predictable ways.

Because the movement is encoded in molecular orientation within the material, LCE systems can produce controlled deformation without mechanical components. Experimental prototypes have demonstrated soft robotic structures capable of locomotion and controlled shape change through photothermal activation. The appeal of LCEs lies in their efficiency and scalability. Instead of translating rotational motor output into linear motion through mechanical linkages, the material converts energy directly into movement. The resulting structures are lightweight, flexible, and capable of continuous deformation. These properties have made liquid crystal elastomers an important focus of research into untethered soft robots and adaptive structures that operate without rigid frames.

Morphing Materials and Fabricated Transformation
Soft robotics is not confined to robotics laboratories. Design researchers increasingly explore how fabrication techniques can encode transformation directly into materials. At the Morphing Matter Lab, directed by Lining Yao, research focuses on materials that change shape in response to environmental stimuli such as heat, moisture, or hydration. Projects from the lab investigate composite structures in which different material layers expand or contract at different rates. When exposed to a trigger, the resulting differential expansion causes bending, folding, or twisting. Examples include hydrogel composites that swell when exposed to water and thermally responsive materials that curve when heated. Because these transformations are defined by material composition and fabrication patterns, no electronic control system is required. The resulting behavior is effectively programmed during fabrication.

This approach reframes fabrication as a form of programming. The geometry of layers, the orientation of fibers, and the distribution of material properties determine how the structure will respond when environmental conditions change.
Such systems have implications across product design, architecture, and robotics, where deployable or adaptive structures can emerge from material behavior rather than mechanical assemblies.

Soft Machines and Fluidic Logic
While some soft robotic systems rely purely on material deformation, others integrate flexible control architectures within the material itself. One example is the use of fluidic circuits—networks of microchannels embedded in elastomers that direct air or liquid pressure through a structure. Research at the Organic Robotics Lab at Cornell University, led by Rob Shepherd, explores soft robots that incorporate pneumatic networks and fluidic logic. In these systems, pressure changes propagating through microfluidic channels can trigger sequences of motion without electronic processors.

The approach builds on earlier research in fluidic computing, where air pressure rather than electrical signals performs logical operations. By embedding these channels directly into silicone structures, engineers can construct robots capable of simple control behaviors such as switching between movement patterns when encountering obstacles. The result is a hybrid system in which sensing, actuation, and control are integrated into a single flexible structure. Instead of assembling motors, sensors, and circuit boards into a rigid machine, the entire system can be fabricated as a continuous soft structure.

Soft robotics research remains a rapidly evolving field, but its broader implications are already visible. Across research labs and design studios, the boundary between materials and machines is beginning to dissolve. Motion can emerge from polymer alignment, magnetic particle distribution, or fluidic pressure networks. Behavior can be encoded through fabrication rather than software. As these approaches mature, the design of mechanical systems may increasingly resemble the design of materials themselves. Movement, sensing, and control become properties of matter rather than functions added through mechanical infrastructure. In that shift, machines begin to look less like assemblies of parts and more like responsive material systems.

The U.S. Supreme Court has declined to review a closely watched case about AI-generated art copyright, leaving intact lower-court rulings that reaffirm a central principle of U.S. copyright law: authorship requires a human creator.

The case, Thaler v. Perlmutter, began when computer scientist Stephen Thaler attempted to register copyright for an image titled A Recent Entrance to Paradise. Thaler stated that the work was produced autonomously by his artificial intelligence system, DABUS, without human involvement. The U.S. Copyright Office rejected the application, concluding that copyright protection applies only to works created by human authors.

Thaler challenged that decision in federal court. In 2023, the U.S. District Court for the District of Columbia upheld the Copyright Office’s determination that the Copyright Act requires human authorship. The U.S. Court of Appeals for the D.C. Circuit later affirmed the ruling. In March 2026, the Supreme Court declined to hear Thaler’s appeal, allowing those decisions to stand.

The outcome reinforces guidance the Copyright Office has issued in recent years: generative AI systems can be used within creative workflows, but works produced entirely without human creative input are not eligible for copyright protection. The ruling does not address another major legal question surrounding generative media—whether AI systems can legally train on copyrighted artworks. That issue is currently being tested in several federal lawsuits involving artists, image libraries, and developers of generative AI systems.

The rapid spread of open-source generative AI models is beginning to reshape how creative work is produced. Instead of relying exclusively on proprietary platforms, a growing number of artists, designers, and researchers are assembling their own machine-learning workflows using openly distributed models and software frameworks. The shift has less to do with the novelty of AI-generated images than with a deeper transformation in the technical infrastructure of creative practice.

Much of this change traces back to the release of Stable Diffusion in 2022 by Stability AI and collaborators. Unlike many earlier generative models, Stable Diffusion was released with publicly accessible weights that allowed developers and researchers to run the model locally and modify its architecture. Within months, an extensive ecosystem of tools and interfaces had emerged around it, turning diffusion-based image generation into a flexible platform for experimentation. Today that ecosystem includes a wide range of software environments that allow users to construct modular AI pipelines, modify training datasets, and fine-tune generative models. What began as a technical release has contributed to a broader shift in how machine learning is integrated into creative workflows.

The Stable Diffusion Ecosystem
Stable Diffusion belongs to a class of generative systems known as diffusion models. These models learn to reconstruct images from noise through an iterative denoising process, gradually producing coherent visual structures. Because the model weights and code were released publicly, developers rapidly began adapting the system to different creative and technical uses.

One important piece of infrastructure supporting this experimentation is Hugging Face, a widely used platform that hosts thousands of machine-learning models and datasets. Through its model hub, researchers and developers can share modified versions of diffusion models, training checkpoints, and experimental architectures. Artists and independent studios frequently download these models directly and run them on local hardware, allowing them to experiment outside commercial AI services.

Alongside the models themselves, a growing set of interfaces has appeared that enable users to construct complex generative pipelines. Software environments such as ComfyUI, Automatic1111’s Stable Diffusion WebUI, and InvokeAI allow creators to build node-based workflows in which image generation can be combined with segmentation models, image conditioning systems, or post-processing stages. In these environments, the generative model becomes one component within a larger computational system. The result is a technical ecosystem in which machine learning operates less like a standalone application and more like programmable creative infrastructure.

Building Modular Generative Pipelines
Node-based AI interfaces have become particularly influential because they allow users to construct visual workflows that resemble production pipelines used in animation or visual effects. Instead of issuing a single prompt to generate an image, users can chain together multiple processes: generating an image, modifying it through image-to-image diffusion, guiding the structure with pose estimation, or altering color and composition through additional model passes. One widely adopted extension of diffusion models is ControlNet, introduced in 2023 by Lvmin Zhang and collaborators. ControlNet allows diffusion systems to follow structural guidance derived from inputs such as edge maps, depth maps, or human pose estimation. This capability gives artists much finer control over the layout and composition of generated imagery.

In practical terms, a generative workflow might involve generating an initial image with a diffusion model, applying ControlNet to enforce compositional constraints, and then refining the output through multiple iterations of image-to-image transformation. Each stage of the process can be modified or replaced, turning the workflow itself into a configurable system. Because these pipelines are modular, they can be adapted to a wide range of creative contexts. Some practitioners use them to produce speculative architectural imagery, while others experiment with animation, motion graphics, or generative illustration. In each case the emphasis shifts away from isolated outputs and toward the design of computational processes.

Running AI Models Inside the Studio
Another significant development is the increasing number of creative studios running generative models locally. Advances in consumer graphics processing units (GPUs) have made it feasible for independent practitioners to run diffusion models on workstation hardware rather than relying entirely on cloud-based services. Running models locally offers several advantages. It allows users to modify models, test experimental architectures, and fine-tune systems using custom datasets. It also makes it possible to construct workflows that combine multiple machine-learning models in ways that would be difficult to implement within a closed platform.

Repositories hosted on Hugging Face distribute models in formats that can be downloaded and integrated into these local pipelines. Developers frequently publish checkpoints for fine-tuned models, allowing others to build upon previous work. The resulting ecosystem resembles open-source software communities, where tools evolve through iterative experimentation rather than centralized development. This distributed infrastructure has accelerated how quickly generative tools are adapted to specific creative needs. Rather than waiting for features to appear in commercial software, practitioners can modify the underlying systems themselves.

From Tools to Infrastructure
The rapid spread of open diffusion models suggests that generative AI may develop less like a single application and more like a layered technological infrastructure. At one level are the base models that generate images. Above them sit the interfaces that allow users to construct pipelines. Surrounding both are repositories, datasets, and software libraries that enable experimentation.

As these layers evolve, the role of creative practitioners is expanding. An increasing number of practitioners now work not only with images or visual compositions, but also with training datasets, generative pipelines, and machine-learning architectures. In this context, the studio begins to resemble a small research environment where models are tested, modified, and integrated into broader production systems.

The emergence of open generative AI ecosystems has not eliminated proprietary tools or platforms. Commercial AI services continue to dominate mainstream use. Yet the rapid development of open models has created an alternative pathway—one in which machine learning can be adapted, modified, and embedded directly into creative workflows. For those building with these systems, the most significant shift may not be the images produced by AI, but the growing ability to shape the tools that produce them.

Kokoska’s compositions frequently reference astronomical imagery, fluid geometries, and atmospheric textures. These elements are combined through digital compositing and AI-assisted image generation, creating compositions that appear simultaneously painterly and computational. The resulting works sit at the intersection of digital collage, generative aesthetics, and contemporary abstract painting.

Rather than positioning generative AI as an autonomous creative agent, Kokoska treats it as one component within a broader visual workflow. Traditional sketching, painting, and compositional development are often combined with digital layering and algorithmic image synthesis. This approach reflects a growing pattern in contemporary creative practice where generative tools operate less as replacements for artistic authorship and more as extensions of studio experimentation.

Across her works, spatial depth is constructed through overlapping color gradients, fragmented shapes, and subtle figurative cues. The compositions often evoke landscapes that appear both planetary and psychological, suggesting environments that exist somewhere between external cosmos and internal perception. These visual strategies allow the work to oscillate between representation and abstraction. Kokoska’s practice is informed by an interest in meditation, consciousness, and altered states of awareness. Rather than depicting these themes directly, the work explores how visual structure—color transitions, motion, and compositional layering—can evoke contemplative attention. Animation and motion graphics further extend this dynamic, introducing temporal shifts that encourage viewers to engage with the imagery gradually rather than as a single static moment.

This emphasis on reflective viewing aligns with a broader shift within digital art toward slower, more immersive visual experiences. While many generative artworks foreground technical spectacle or algorithmic complexity, Kokoska’s pieces instead focus on atmospheric depth and perceptual ambiguity. The work invites viewers to spend time within the image, allowing subtle relationships between form, movement, and color to emerge.

Kokoska’s background in graphic art and visual communication informs the clarity of her compositional structure. She holds a Master’s degree in Graphic Art and studied at the Toronto School of Art, alongside independent studies in Bali. Her practice has developed through a combination of formal education and extensive self-directed experimentation across both physical and digital media.

Through Infinite Mantra Art Studio, Kokoska continues to explore the evolving relationship between digital tools and contemplative visual culture. Her work reflects a wider moment in contemporary creative practice where artists are integrating generative systems into hybrid workflows that combine traditional studio methods with computational image production.

At Paris Fashion Week, ANREALAGE’s Fall/Winter 2026 collection introduced garments embedded with LED displays designed to visually blend into their surroundings. Designer Kunihiko Morinaga presented coats, dresses, and structured outerwear that reproduced fragments of the projected runway environment, creating a camouflage-like effect as models moved through the space. The garments functioned as programmable visual surfaces, displaying imagery synchronized with the projections surrounding the catwalk. The concept draws on thermoptic camouflage, a fictional technology depicted in Ghost in the Shell, the cyberpunk manga created by Masamune Shirow.

Clothing as a Digital Display
The FW26 collection centered on garments that functioned more like modular display surfaces than traditional textiles. Coats, dresses, and structured outerwear incorporated LED panels capable of presenting animated imagery. During the show, large-scale projections formed the visual backdrop of the runway while the garments displayed fragments of the same visuals through embedded displays.

When alignment occurred between projection and garment, sections of the clothing visually blended with the background imagery, producing the illusion that parts of the wearer were dissolving into the environment. The effect relied on coordination between stage design and garment programming. LED displays embedded within the clothing rendered digital imagery synchronized with the projected visuals surrounding the runway.

According to the brand, Morinaga and his team spent several months developing the programming system used in the show. The process involved calibrating LED panels, synchronizing visuals with runway projections, and ensuring the garments remained structurally wearable. This differs from many projection-mapping runway presentations, where garments primarily serve as surfaces for projected imagery. In the ANREALAGE show, the garments themselves displayed digital visuals through embedded LED panels synchronized with the surrounding projections.

Cyberpunk References in Material Form
While the LED garments formed the technical core of the collection, the surrounding designs incorporated recognizable clothing elements. Prairie-style shirts, fringed denim, and softer fabrics appeared beneath or alongside the illuminated structures. This juxtaposition created a contrast between programmable surfaces and more conventional garment forms. Digital display components were layered over familiar silhouettes, combining electronic materials with traditional clothing construction.

Several looks incorporated structured, armor-like shapes alongside bright neon color palettes. Metallic accessories and hardware-like details appeared throughout the collection, reinforcing the technological aesthetic established by the LED garments. The collection also referenced thermoptic camouflage from Ghost in the Shell. The cyberpunk narrative imagines surfaces capable of reproducing their surroundings in order to conceal objects, a concept Morinaga translated into wearable display garments synchronized with the runway projections.

Fashion as a Testing Ground for Emerging Materials
Morinaga founded ANREALAGE in 2003 and has frequently incorporated experimental materials and technologies into his collections. Previous runway presentations have included ultraviolet-reactive fabrics, garments that change color under different lighting conditions, and pieces designed for projection mapping.

Many of these projects explore how clothing surfaces respond to light, projection, and environmental inputs. The FW26 runway extends this approach by integrating LED display systems directly into garments. While wearable LED technology has appeared in experimental fashion contexts before, the ANREALAGE presentation incorporated these displays into the structure of the runway environment itself. Background projections generated the visual environment while the garments displayed synchronized digital imagery.

Runway presentations provide a controlled setting where designers can experiment with new materials, visual systems, and interactions between clothing and digital media. In the FW26 show, thermoptic camouflage was explored through garments equipped with LED displays synchronized with the projections surrounding the runway.

Brain–computer interfaces (BCIs) have traditionally been developed within medical and neuroscience research. Systems capable of translating electrical activity in the brain into machine-readable signals have enabled patients to control prosthetic limbs, communication interfaces, and computer cursors through neural activity. Over the past decade, artists and designers have also begun experimenting with these technologies, exploring how neural signals can function as expressive inputs rather than purely clinical data.

Advances in affordable electroencephalography (EEG) headsets and open-source signal processing frameworks have made neural sensing increasingly accessible outside laboratories. Although EEG devices cannot decode thoughts, they can detect measurable patterns associated with attention, relaxation, and cognitive effort. Within creative practice, these signals are often treated not as precise commands but as generative inputs capable of influencing visual systems, sound synthesis, or responsive environments.

Across contemporary media art, neural interfaces are emerging as instruments that translate cognitive signals into visual, sonic, and spatial media. Early experiments in biofeedback art during the 1960s and 1970s similarly explored how physiological signals—heart rate, brainwaves, or skin conductance—could be translated into audiovisual systems. Today’s BCI-based artworks extend this lineage, combining neural sensing with machine learning, real-time computation, and networked media environments.

EEG Signals as Visual and Physical Media
One of the most direct artistic uses of EEG appears in the work of artist Lisa Park, whose installations translate brain activity into physical phenomena. Park frequently uses consumer EEG headsets to capture neural signals associated with shifts in attention and cognitive engagement.

In the performance installation Eunoia (2013), Park wears an EEG device while seated among a circular array of water-filled vessels. Each vessel is connected to a vibration transducer that responds to changes in the EEG signal. As the performer attempts to regulate her mental focus, the vibrations create visible ripples across the water’s surface. The system does not attempt to interpret specific thoughts. Instead, it converts fluctuations in neural activity into subtle physical transformations. The resulting patterns make cognitive dynamics visible, translating otherwise imperceptible neural activity into an observable environmental effect.

Park’s work demonstrates how neural interfaces can function less as instruments of control than as mechanisms for externalizing cognition. By turning EEG signals into physical motion, the installation foregrounds the fluctuating nature of neural activity while revealing how internal mental states can shape responsive systems.
Park extended this approach in the installation Blooming (2015), which translates EEG signals into kinetic movement within a spatial environment. In the work, an EEG headset captures fluctuations in neural activity associated with attention and mental engagement. These signals are transmitted to mechanical devices that animate artificial flowers placed throughout the installation.

As the participant’s brain activity shifts, the flowers open and close in response to variations in the EEG data. The installation creates a feedback loop between internal cognitive states and the surrounding environment. Small changes in attention or mental effort become visible through the movement of the flowers. Like Eunoia, Blooming does not attempt to decode thoughts. Instead, the system translates measurable neural patterns into a responsive physical system. The work highlights the instability and interpretive nature of neural signals while demonstrating how cognitive data can be transformed into expressive media.

Neural Signals in Computational Media
Artist and technologist Daito Manabe, founder of Rhizomatiks, investigates how neuroscience research and machine learning intersect with computational media. His work often examines how physiological and perceptual data can be translated into visual systems. Dissonant Imaginary (2019), developed by Rhizomatiks Research in collaboration with neuroscientist Yukiyasu Kamitani, draws on Kamitani’s research into neural decoding, where machine-learning models analyze patterns of brain activity associated with visual perception.

Rather than functioning as a conventional brain–computer interface, the work explores how neural data can be computationally interpreted and represented. Visual systems developed for the project reference research that reconstructs or predicts images from brain activity, highlighting how machine-learning models translate neural signals into visual information. By connecting neuroscience research with generative media systems, the project reflects the growing intersection between brain research and experimental media practice.

Human–Machine Perception and Cognitive Systems
Artist Justine Emard investigates how artificial intelligence and sensing technologies interpret human presence. Her installations frequently stage encounters between human bodies and intelligent systems capable of analyzing movement, gesture, and behavioral patterns.

In Co(AI)xistence (2017), developed with researchers associated with the Okinawa Institute of Science and Technology, Emard presents an interaction between a human dancer and a robotic arm equipped with AI-based vision systems. The system observes the dancer’s gestures through cameras and computational sensing technologies, generating responsive movements that unfold throughout the performance.

Rather than translating neural signals directly into media output, the project explores how machines construct representations of human behavior through sensor data and machine learning models. The installation highlights the interpretive frameworks through which computational systems perceive and model human activity. By situating the body within networks of sensors and intelligent systems, Emard’s work reflects the broader technological landscape in which neural interfaces are emerging. Brain signals represent only one layer within a growing ecosystem of biometric data streams captured and interpreted by computational systems.

Across these practices, neural interfaces appear less as tools for reading minds than as instruments that translate cognition into data streams capable of shaping media systems. The result is not mind-reading but a new class of responsive technologies in which attention, perception, and mental effort become parameters within computational environments.

A comprehensive political and design theory of planetary-scale computation proposing that The Stack—an accidental megastructure—is both a technological apparatus and a model for a new geopolitical architecture.

What has planetary-scale computation done to our geopolitical realities? It takes different forms at different scales—from energy and mineral sourcing and subterranean cloud infrastructure to urban software and massive universal addressing systems; from interfaces drawn by the augmentation of the hand and eye to users identified by self—quantification and the arrival of legions of sensors, algorithms, and robots. Together, how do these distort and deform modern political geographies and produce new territories in their own image?

In The Stack, Benjamin Bratton proposes that these different genres of computation—smart grids, cloud platforms, mobile apps, smart cities, the Internet of Things, automation—can be seen not as so many species evolving on their own, but as forming a coherent whole: an accidental megastructure called The Stack that is both a computational apparatus and a new governing architecture. We are inside The Stack and it is inside of us.

In an account that is both theoretical and technical, drawing on political philosophy, architectural theory, and software studies, Bratton explores six layers of The Stack: Earth, Cloud, City, Address, Interface, User. Each is mapped on its own terms and understood as a component within the larger whole built from hard and soft systems intermingling—not only computational forms but also social, human, and physical forces. This model, informed by the logic of the multilayered structure of protocol “stacks,” in which network technologies operate within a modular and vertical order, offers a comprehensive image of our emerging infrastructure and a platform for its ongoing reinvention.

The Stack is an interdisciplinary design brief for a new geopolitics that works with and for planetary-scale computation. Interweaving the continental, urban, and perceptual scales, it shows how we can better build, dwell within, communicate with, and govern our worlds.

thestack.org

Nxt Museum’s Still Processing is a group exhibition that interrogates the mechanics of perception in a mediated world — a space where datasets and algorithms no longer sit behind screens but directly shape what we see, hear, and feel. Curated by Belgrade-born Bogomir Doringer and on view at Nxt Museum in Amsterdam from 7 February to 5 October 2025, the exhibition assembles seven contemporary artists whose works make tangible the often invisible processes of image generation, transformation, and interpretation.

At its core, Still Processing operates with a dual premise: first, that technological systems — from compression algorithms to AI synthesizers — materially shape the images and sounds circulating in our culture; second, that human sensory mechanisms themselves are active processors of information, constructing meaning through a complex interplay of cognition, expectation, and sensation.

The Exhibition’s Structural Logic: Algorithm and Embodiment
Rather than presenting a linear narrative, Still Processing is organized around staged encounters with work that foregrounds computational logic as a shaping force. Across seven large installations and transitional spaces, visitors move through perceptual registers where light, sound, motion, and digital representation are not just exhibited but activated.

A foundational strand in the show is Rosa Menkman’s suite of research-based works, which trace pivotal moments in the history of image technologies. Spanning pieces such as The Collapse of PAL and IM/POSSIBLE RAINBOWS, Menkman’s practice unpacks how standards, compression, and visual protocols influence what is preserved — and what is lost — when images are encoded, transmitted, and decoded digitally. These works perform a kind of archaeology of visual media, revealing the compromises and hidden biases that underlie seemingly neutral technical systems.

The exhibition also places computational aesthetics in dialogue with the body’s physical reception of stimuli. Gabey Tjon a Tham’s Red Horizon uses kinetic pendulums and light to generate patterns that produce afterimages on the viewer’s retina — a direct intersection of mechanical motion and human visual memory. Lumus Instruments’ Polynode XI similarly translates audiovisual data into an ambient field that continually evolves in response to environmental cues, collapsing the subject/object distinction between viewer and machine signal.

Between Synthetic Worlds and Sensory Conditions
In several works, the technologies of image and sound generation are more than instruments — they become subjects of inquiry. Geoffrey Lillemon’s Simulation in Blue situates AI-generated figures in an evolving audiovisual environment, where CGI and algorithmic choreography produce a hybrid form that defies conventional categorization.

Using tools associated with both entertainment and generative art, the piece raises questions about authorship and agency in synthetic media.
A contrasting spatial intervention comes from Children of the Light, whose ALL-TOGETHER-NOW creates a constellation of floating luminous rings inspired by the first iconic image of a black hole. Here, the interplay of light and spatial perception foregrounds the limits of visualization — especially when confronting phenomena that sit at the margins of human sensory capacities.

The exhibition’s commissioned project by Balfua, The Slollaleia, occupies a dedicated stage in the museum, deploying AI and digital world-building to evoke a speculative ecology of synthetic beings. Drawing on animation and algorithmic behaviour, this work positions machine logic not as an external force but as a generative partner in the creation of new forms.

Finally, Boris Acket’s Duration uses sound and light in a room-scaled structure that questions our experience of time itself. Instead of linear sequencing, the installation generates patterns that expand and collapse in real time, revealing how perceptual time — like image processing — is contingent, constructed, and continually re-negotiated.

Processing as Reflection, Not Metaphor
What distinguishes Still Processing from more typical tech-centric exhibitions is its insistence on processing as an active condition rather than a metaphor. The works do not simply illustrate technological influence; they materialize the act of translation between encoded signals and human interpretation. This emphasis on material processes — whether through algorithmic generation, kinetic motion, or sensory feedback — foregrounds the active, iterative nature of perception itself.
Across the exhibition’s trajectory, visitors can perceive how machine logic does not operate in isolation: it infiltrates standards, expectations, and even cognitive habits. Likewise, the human nervous system is shown not as a passive receptor but as part of a continuous loop of interpretation and re-interpretation. 

Still Processing thus situates the question of meaning not in static representation but in ongoing coupling between human systems and technological media. While the works diverge in media and mode — from archival probes into compression histories to speculative simulations of imaginary ecosystems — they share a focus on the seams where processing becomes visible, where data ceases to be inert and begins to shape experience. For a contemporary audience attuned to the ubiquity of digital mediation, the exhibition provides frameworks for understanding how meaning remains contingent, negotiated, and still very much under construction.

The stated premise was straightforward: create space to experiment. Each contributor worked within their respective discipline—Bosmans on design and illustration, Waltz on animation and project direction, and Poon on sound design—while responding to the evolving output of the others. The result is a tightly integrated audiovisual work shaped through reciprocal exchange rather than a fixed production hierarchy.

The collaboration was presented at the OOO Exhibition in Oslo in 2023 and is scheduled to be shown at the Northbridge Piazza Screen in 2024, positioning the work within both exhibition and public-screen contexts.

The project reflects a contemporary motion practice in which plugins, animation systems, and sound design tools operate as shared infrastructure across disciplines, supporting distributed authorship rather than isolated production roles.

Project Info
Year: 2023
Project Lead & Animation: Anders Waltz
Design & Illustration: Jerry-Lee Bosmans
Sound Design: John Poon
Tool: Flex plugin (aescripts)
Presented at: OOO Exhibition, Oslo (2023)
Screening: Northbridge Piazza Screen, Perth (2024) Image Credit: Jerry-Lee Bosmans, Anders Waltz, John Poon

Fluid Current consists of water, aluminium, an electric circuit, and an LED. As water moves through the aluminium structure, it completes an electrical circuit and triggers a pulse of light. The system is direct: conductivity produces illumination.

The work focuses on the relationship between material behavior and perceptual experience. Water—commonly understood as fluid and unstable—functions here as a conductive element within a simple electrical system. The circuit is not concealed; its operation unfolds physically through contact and flow.

Hagelaar has described his practice as engaging two related perceptual states: estrangement, when familiar materials reveal unexpected properties, and stilling, when subtle movement concentrates attention. In Fluid Current, both are present. The timing and intensity of the light depend on the dynamics of the water, introducing variation into an otherwise minimal setup.

By aligning a natural process with a technical response, Fluid Current frames perception as an encounter with cause and effect in real time. The piece does not simulate interaction; it stages it through material conditions.

Project Info
Artist: Sander Hagelaar
Materials: Water, Aluminium, Electric Circuit, LED

Year: 2020

Slow Technology Reader gathers contributions from diverse disciplinary fields and knowledge traditions to consider technology through a ‘Slow’ lens. By turns artistic, speculative, and academic the contents here probe alternative potentials for the digital entities proliferating in our midst, by invoking variable pacings and temporalities of engagement; reflecting through tools and techniques that have endured the test of time; and looking to non-Western and more-than-human sources to inspire technological development.

This new volume in the Slow Reader series gestures toward a fuller spectrum of what technology is and can be, moving beyond the limited perspectives and legacy structures that dominate technological development today. It includes the rich insights and intelligences of feminist, queer, Indigenous, activist, and ecological practices—offering them as vibrant data points for shaping more just and generative futures. At a time when the digital reaches into nearly every facet of planetary existence, this book aims to disrupt and recalibrate how we think about and relate/live with technology, illuminating more expansive pathways forward.

Contributions: Paula Albuquerque, Kader Attia, Aïsta Bah, Scott Benesiinaabandan, Pac.me B.ru, Cláudio Bueno, Derrais Carter, Raven Chacon, Joana Chicau, Guy Cools, Laura Coombs, Siobhán K. Cronin, Will Daddario, Edwidge Danticat, Thierno Dia, Mamadou Taslim Diallo, Henriette Essami-Khaullot, Silvia Federici, Mariana Fernández Mora, Ella Finer, Jem Finer, Mashinka Firunts Hakopian, Dakin Hart, Faïza Hirach, Candice Hopkins, Christine Hvidt, Carol R. Kallend, Theun Karelse, Danel Khojayeva, Suzanne Kite, Fran Kourouma, Jaron Lanier, Jason Edward Lewis, Pia Lindman, Gļeb(s) Maiboroda, Pierre Marchand, Michael Marder, Nanako Nakajima, Florence Okoye, Marina Orlova, Jogi Panghaal, Moisés Patrício, Rory Pilgrim, Elisabeth (elieli) Raymond, Milady Renoir, Oscar Santillán, Laurel Schwulst, Mindy Seu, Camila Sposati, Christel Stalpaert, Corey Stover, Melita Stover Janis, Carolyn F. Strauss, Foluke Taylor, Alberto Isifin Tchama, Ovidiu Ţichindeleanu, Rolando Vázquez Melken, Evelyn Wan, Halidou Wuandaougo, Arkadi Zaides, Joanne Zerdy, Martín Zícari

Support: Creative Industries Fund NL, Cultuurfonds

Design: Haller Brun

Real-time systems promise immediacy. In AI interfaces, gaming engines, financial markets, and live media, lower latency is equated with intelligence, efficiency, and control. Response time becomes a proxy for authority: the system that answers fastest appears most capable.

Latency, however, is not simply a technical metric. It is a distribution of responsiveness. It determines who speaks, who waits, and what counts as recognition. In contemporary performance and computational art, delay becomes material—staged, modulated, and made perceptible. Rather than treating latency as error, artists work with it as structure.

Across theater, immersive environments, and live electronic performance, Annie Dorsen, OpenEndedGroup, and Myriam Bleau and Myriam Bleau demonstrate how real-time systems redistribute agency through computational mediation.

Latency as Dramaturgy
Annie Dorsen’s theater projects foreground computation in real time. In Hello Hi There (2010), a live performance structured around algorithmically recombined philosophical debate, text is generated and delivered on stage through software systems drawing from pre-authored corpora. In The Slow Room (2018), Dorsen employed machine-learning systems trained on canonical dramatic texts to generate dialogue live, shaping each performance through probabilistic output rather than fixed script.

More recently, in Yesterday Tomorrow (2023), Dorsen incorporated large language models into a live performance structure, allowing text to be generated dynamically rather than performed from a predetermined script. The work extends her long-standing investigation of machine-authored speech as a theatrical condition.

Across these projects, language unfolds through computational processing rather than predetermined sequence. The audience witnesses dialogue emerging from a system in operation. Pauses between exchanges are structured not only by performers but by the processing dynamics of the software itself. Speech is selected, assembled, and delivered in real time.

Timing shifts authority on stage. The performer does not fully determine pacing; system response participates in shaping rhythm and tension. Silence becomes infrastructural rather than purely psychological. When language is generated live, the interval between prompt and output becomes part of the performance’s architecture. Consumer AI platforms are engineered to minimize visible delay in order to simulate fluency. Dorsen’s work instead keeps mediation legible. Response time is shown to be conditional and shared. Latency becomes dramaturgy.

Latency as Environmental Intelligence
If Dorsen stages latency as dialogue, OpenEndedGroup spatializes it. Founded by Marc Downie and Paul Kaiser, the collective works at the intersection of dance, computation, and immersive projection. Their installations and performances employ motion capture and real-time rendering systems that interpret human movement and generate visual response.

Gesture is captured, translated algorithmically, and returned as light, image, or spatial transformation. The response unfolds in near real time, though never without mediation. Action and output are coupled, but not fused. Tracking systems operate through predefined thresholds and parameters. Recognition depends on configuration. What appears seamless is structured by computational limits and decision rules.

This logic extends beyond the gallery. In biometric authentication, autonomous vehicles, and surveillance systems, response time governs classification and access. Speed often signals certainty; delay can register as ambiguity. OpenEndedGroup’s environments render that structure perceptible. The space responds—but on the basis of system interpretation. Latency here is not spectacle. It is the condition of interaction..

Latency as Kinesthetic Feedback
Myriam Bleau’s performance systems make computational response tangible. In Soft Revolvers (2014), she performs with a set of translucent, illuminated discs that function as both instrument and interface. Each object contains embedded sensors that track rotation and gesture, transmitting data to custom software that generates and modulates sound in real time.

Movement is translated into signal. Signal becomes audio output. The response is immediate but mediated through software interpretation. Unlike acoustic instruments, where vibration directly produces sound, Bleau’s system depends on computational mapping. Gesture must be read, processed, and rendered.

That interpretive layer structures timing. The performer adjusts not only to rhythm and composition, but to the responsiveness of the system itself. Expressivity emerges from the interplay between body and computational response.

In Second Self (2024), created with Nien Tzu Weng, Bleau extends this exploration of responsive systems through portable interactive LED screens. Inspired by Sherry Turkle’s concept of the “second self,” the work treats the interface as a bodily extension. Each gesture alters sound and visual output in real time, staging an audiovisual dialogue between human movement and technological mediation.

Here, latency is not framed as delay but as the condition of exchange. Timing shapes interaction, and interaction shapes authority.

Timing as Power
Across these practices, latency is not inefficiency. It is structure. In technical discourse, low latency is framed as progress. Edge computing reduces network delay by bringing computation closer to users. High-frequency trading firms compete on microseconds. AI inference optimization accelerates model response times. In many contexts, faster systems gain competitive advantage. Performance and computational installation complicate that narrative. Rather than eliminating delay, they foreground mediation. Response time becomes visible. Interaction is shown to be organized by timing.

Who receives immediate feedback? Who waits? Who determines the pace of exchange?

By foregrounding computational timing, Dorsen, OpenEndedGroup, and Bleau shift attention from output to responsiveness. The encounter becomes a study of how systems listen, process, and reply. Latency shapes rhythm, choreography, and perception. It also influences how authority is read. As real-time AI increasingly mediates speech, movement, and image, timing extends beyond performance. Every “real-time” interaction is structured by decisions about speed.

Each page is 3D printed with conductive TPU traces positioned beneath selected lines of text. These traces connect to capacitive touch sensors, allowing readers to activate audio output through direct contact with the printed surface. A microcontroller, speaker module, and electronic ink display are housed within a steel-infused PLA spine and cover, consolidating the computational components within the book’s physical architecture. The circuitry is not attached after printing; it is embedded into the page structure itself, making the book both printed artifact and functional device.

The project takes conceptual direction from Ray Bradbury’s short story “Ylla,” from The Martian Chronicles, which describes a metal book inscribed with raised hieroglyphs that emit sound when brushed by hand. METALLY interprets this description as a fabrication premise. The layered relief inherent to extrusion-based 3D printing—raised letterforms, surface ridges, volumetric marks—becomes both typographic strategy and tactile interface. In this configuration, touch is not symbolic; it is operational.

The development of METALLY builds on several years of experimental publishing at Hyper Press. Earlier research into flexible 3D printed substrates made it possible to produce page-like forms capable of bending while retaining embedded conductive pathways. Custom typefaces were designed specifically for extrusion-based production using a methodology termed path-trapping, which stabilizes letterforms during printing. Multi-material mechanical interlocking techniques, referred to as clotting, were developed to allow different printed materials to bind structurally without adhesives. These fabrication strategies converge in METALLY as a fully integrated workflow in which text, image, substrate, and electronics are co-produced.

By embedding circuitry directly into printed matter, the project situates 3D printed books within a longer lineage of press technologies and graphic design. It does not frame additive manufacturing as a departure from print history, but as an extension of it—revisiting typesetting, plate-making, and binding through digital fabrication and computational control. In doing so, the book expands beyond a purely visual medium. The page registers touch, produces sound, and incorporates dynamic display, introducing additional sensory and technical layers while remaining legible as a book-object.

METALLY (2025) was selected as Excellent Work in the Experimental Work category for the Tokyo Type Directors Club Annual Award 2026. The recognition positions the project within contemporary graphic design discourse, particularly in relation to expanded definitions of typography, material production, and the future of the printed page.

Project Info
Project: @darius_ou
Publisher & Research Platform: @hyper.press
Support: @feelers_feelers, @joandkapi
Guidance: @hiddnur, @sojamo, @msjospark, Dr. Aprille
Photography: @jonathantyl
Talent: @suvalidh

Generative AI is typically experienced through screens: chat windows, image feeds, dashboards. Increasingly, though, machine learning systems are being embedded into physical objects—arcade cabinets, kinetic installations, responsive environments—where computation unfolds through motors, water, light, and mechanical movement rather than pixels.

This shift does not reject software. It relocates it. The interface is no longer only a touchscreen but a surface, a sensor array, a hinge, a grid of valves. The output is not just a rendered image but a physical change in space. The move from screen to object becomes visible in the work of Ross Goodwin, Random International, and Universal Everything.
 
Ross Goodwin: Machine Learning as Physical Apparatus
Ross Goodwin is known for experiments in computational writing. In 2017, he created 1 the Road, a book generated during a cross-country car trip using sensors connected to a recurrent neural network. Cameras, GPS, microphones, and a clock supplied real-time data to the system, which produced text continuously and printed it on receipt paper as the journey unfolded.

The setup occupied a passenger seat. It required cameras, wiring, computing hardware, and a thermal printer. Text was generated and printed live during the trip, foregrounding the mechanics of machine inference rather than hiding them behind an interface. The published book was compiled from that output. Goodwin has also exhibited installations in which AI-generated text is printed in real time, making the act of computation materially visible. Paper feeds. Ink accumulates. Lines appear sequentially. Instead of interacting with a seamless interface, viewers encounter a machine performing inference in physical space.
Here, AI is not a background service. It is equipment.

Random International: Responsive Environments
The London- and Berlin-based studio Random International has long worked with robotics, sensing, and responsive systems. Their still-infamous project from 2012, Rain Room, uses motion tracking and a ceiling-mounted grid of electronically controlled valves to regulate falling water. Visitors walk through a downpour that parts around their bodies in real time. Rain Room, is AI, but its operation depends on real-time sensing and control systems. Cameras detect a visitor’s position; software calculates a response; hardware actuates a dense grid of individually controlled water streams. The result is an environment that reacts to human movement without relying on screens or projections.

Subsequent projects extend this approach. Random International has continued to develop responsive, system-driven installations. In Life in our Minds: Motherflock (2023), a coordinated group of robotic light elements moves collectively in response to real-time behavioral logic. Individual units operate as part of a distributed system, adjusting orientation and intensity according to programmed rules and environmental input. Computation is expressed through motors, control systems, and structural assemblies rather than screens.

Interaction remains spatial and bodily. Viewers enter the environment, and the system responds through coordinated physical motion. The technology is embedded in infrastructure rather than presented as a graphical interface.

Universal Everything: AI in Motion Systems
The UK-based studio Universal Everything is known for large-scale digital character works and immersive installations. In recent years, the studio has incorporated machine learning into projects where movement and form are generated or influenced algorithmically rather than pre-rendered as fixed sequences. In exhibition and public-facing installations, generative systems determine how digital figures evolve in response to real-time inputs. In some contexts, these outputs extend beyond the screen into architectural lighting systems or sculptural LED arrays, where software directly controls physical illumination patterns.

Instead of playing back completed animation files, systems calculate motion live. Sensors supply data; algorithms determine variation; lighting and display hardware execute change. The display remains, but it functions as part of a broader computational apparatus tied to physical infrastructure.

From Interface to Infrastructure
Across these practices, a pattern emerges. AI is not confined to a graphical user interface. It is integrated into cabinets, rooms, lighting systems, and mechanical assemblies. This approach is also visible in custom-built interactive devices presented at technology conferences and design fairs, where rule-based or generative systems are embedded directly into playable hardware. Buttons trigger algorithms. Joysticks modify system behavior. LED grids visualize live computation. The system is experienced through physical controls rather than abstract menus. Physical interfaces impose constraints. Motors have torque limits. Water systems require calibrated pressure. Printers jam. These conditions shape how generative systems are engineered and deployed. Latency becomes perceptible. Malfunctions appear as mechanical disruptions rather than rendering artifacts.

Re-embedding AI into hardware alters perception. When a language model produces text in a browser, the process remains opaque. When a system prints text onto paper in real time, the sequence becomes legible: sensor input, computation, mechanical output. Cause and effect are easier to trace because the system occupies space. Embedded computation is also present in certain architectural and fabrication contexts. Research studios have developed responsive partitions, adaptive lighting arrays, and sensor-driven prototypes that adjust position or illumination based on environmental input. In these cases, sensing and control systems are integrated directly into structural components rather than mediated solely through apps. What distinguishes these projects is not novelty but placement. The model is not accessed exclusively through a browser window. It is situated within the object.

Generative systems without screens do not eliminate digital media. They redistribute it. Computation moves closer to actuators and sensors. Interfaces become spatial rather than exclusively graphical. The visible output is movement, light, water, or ink. For a field saturated with software demos and chat interfaces, this marks a measurable shift. AI is not only something prompted through text. It is something that occupies space, consumes power, and moves matter.
Not on a screen. In the room.

Rather than generating text or images for a screen, NOEMA segments and reconstructs perception through sound. The system delivers spatialized audio experiences—ambient cues, narrative fragments, and musical elements—that respond to context and reshape how a listener interprets their surroundings. In this configuration, sound is not supplementary; it is the primary interface.

The project reframes AI as an atmospheric layer rather than a discrete output mechanism. Instead of presenting information visually, NOEMA positions machine intelligence as an interpretive presence—an inner voice that describes, questions, and reframes what is sensed. This shift foregrounds listening as an active mode of interaction and suggests alternative models for human–AI engagement rooted in embodiment and spatial awareness.

By relocating computation from screen-based display to acoustic space, NOEMA examines how perception changes when intelligence is encountered as environment rather than image.

Project Info
Developers: Melo Chen and Nomy Yu

Year: 2025
Image Credit: NOEMA by Melo Chen & Nomy Yu

Scale has defined machine learning for more than a decade. Foundation models are trained on scraped web corpora and optimized through centralized cloud infrastructure. Performance is measured in parameters and benchmarks. More data, more compute, more reach. Yet across artistic and research practices, a different logic is emerging—one that treats scale not as destiny but as design.

Dataset minimalism does not simply mean “small models.” It means interrogating how datasets are assembled, what they extract, and what their expansion requires. The politics of AI are embedded not only in model architectures but in the construction of their training ground. The contrast becomes clearer when looking at three distinct but interconnected approaches: Memo Akten’s constrained neural networks, Kate Crawford and Vladan Joler’s infrastructural cartographies, and Joana Moll’s quantification of digital extraction.

Building Within Limits: Memo Akten
In 2017, Memo Akten presented Learning to See, an interactive installation in which a neural network was trained on images captured live through a webcam. Rather than depending on a vast pre-scraped internet corpus, the system learned from a locally generated dataset shaped by participants in the space. Visitors could watch the training process unfold and observe how specific examples influenced the model’s output.

The project made a foundational aspect of machine learning visible: models reflect the structure of their training data. When the corpus is small and legible, causality becomes easier to trace. Misclassification is not an abstract glitch—it reflects what the system has or has not been shown. Akten has spoken about how bias and opacity are embedded in training data. By working with purpose-built datasets, the focus shifts upstream. Instead of treating bias as a downstream bug to patch, dataset construction becomes a primary design concern.

Mapping the Infrastructure of Scale: Crawford and Joler
In Anatomy of an AI System (2018), Kate Crawford and Vladan Joler mapped the supply chains behind a single consumer device: the Amazon Echo. The diagram traced lithium and rare earth mineral extraction, global logistics, data labor, and server infrastructure—demonstrating that AI systems depend on vast material and human networks. Their later project, Calculating Empires (2023), expanded this analysis historically and geopolitically, charting how computational systems have consolidated political and economic power across centuries.

Where Akten makes dataset construction legible at small scale, Crawford and Joler expose what large-scale systems obscure. The project suggests that scaling AI systems entails expanded material, labor, and energy infrastructures. Scale is infrastructural. Seen from this vantage point, dataset minimalism is not aesthetic restraint. It is a structural intervention. Limiting data aggregation limits the expansion of the infrastructures that sustain it. But infrastructure is not only visible in supply chains. It is measurable in energy consumption and tracking systems that operate quietly in the background of everyday digital life.

Measuring Extraction: Joana Moll
Joana Moll’s practice quantifies the hidden material costs of digital systems. In CO2GLE, she calculates the carbon emissions generated by Google searches. In The Hidden Life of an Amazon User (2019), she traced the tracking mechanisms and data flows embedded in a single user session, revealing the complex web of third-party exchanges behind routine interactions.

Moll’s work does not train models; it measures the infrastructures that enable large-scale data capture. By translating digital processes into energy consumption, economic value, and environmental impact, she reframes scale as material burden. As datasets expand, storage and compute demands increase. As models scale, so do the data center infrastructures required to sustain them. The abstraction of “big data” carries measurable physical consequences.

Placed alongside Akten’s constrained training systems and Crawford and Joler’s cartographies, Moll’s quantifications complete the picture. Small datasets are not inherently virtuous, but they are more traceable. Their dependencies are easier to identify.

Designing Against Extractive Scale
Taken together, these practices reposition dataset minimalism as a design stance. Akten demonstrates how working with bounded corpora clarifies authorship and bias. Crawford and Joler reveal that large-scale AI systems depend on expansive material and geopolitical infrastructures. Moll measures the environmental and economic costs embedded in digital accumulation.

The through-line is visibility. At industrial scale, causality disperses across continents and supply chains. At smaller scales, relationships tighten. Decisions about inclusion, labeling, storage, and consent become more explicit. Dataset minimalism does not reject machine learning; it questions the assumption that performance must scale with extraction. It treats scale as optional rather than inevitable. As generative systems proliferate, pressure to expand training corpora will continue. These practices suggest another metric for rigor: not only how well a model performs, but how clearly its dataset can be traced—who contributed to it, what infrastructures sustain it, and what consequences it produces. In that sense, designing with less is not a retreat from AI. It is a demand for accountability in how intelligence is constructed.

High-end laptops equipped with dedicated GPUs and AI accelerators now handle real-time ray tracing, multi-stream 4K editing, parametric modeling, and generative image synthesis without tethering creators to fixed workstations. Platforms such as NVIDIA Studio—and comparable silicon from Apple, Intel, and AMD—are increasingly optimized for creative workloads. This may not simply a performance upgrade – perhaps its an infrastructure shift.

GPU acceleration and on-device AI tooling are compressing workflows that once required towers, external drives, or centralized studios into portable systems. Features such as AI upscaling, generative fill, denoising, background removal, and real-time simulation are increasingly accelerated at the hardware level. Creative software still runs across platforms—but it is increasingly optimized around specific silicon architectures.

As rendering, simulation, and AI inference move onto mobile hardware, more stages of production can occur outside fixed studios. Musicians can travel with synthesis, sampling, and mastering environments in a single device. Designers iterate on complex 3D scenes on site. For some teams, reliance on centralized render infrastructure decreases. The boundary between fieldwork and post-production becomes more fluid.

But mobility also consolidates ecosystems. Hardware acceleration is often aligned with specific software stacks. Drivers, neural engines, and GPU architectures are optimized for particular creative platforms, shaping workflows around tuned pipelines. Portability expands access—while subtly influencing how creative tools are configured and used.

The desktop is not disappearing. High-end studios, large simulations, and heavy training workloads still benefit from fixed infrastructure. But for designers reviewing models on site, musicians producing while touring, or small teams working across cities, this flexibility changes how and where work happens.

Rather than using a keyboard, sequencer, or touch screen, the instrument is played through two handheld electromagnetic pickups. These pickups sense variations in an electromagnetic field generated by the instrument’s surface. As the performer moves their hands through space, shifts in proximity and orientation modulate pitch, timbre, and amplitude.

The system is fully analog and polyphonic. Each hand controls an independent voice, enabling harmonic and intervallic relationships to be shaped through continuous gesture rather than discrete keys. The interface is touch-sensitive in the sense that it responds to embodied motion, but the sound is produced without direct physical contact. The performer navigates an invisible field.

Chromaplane repositions synthesis as a spatial practice. Instead of selecting presets or triggering predefined notes, the musician calibrates tone through subtle changes in distance and angle. This produces a mode of performance that foregrounds proprioception, fine motor control, and sustained attention.

In 2025, Chromaplane received First Prize at the Guthman Musical Instrument Competition, a competition hosted by Georgia Institute of Technology recognizing innovation in new musical instrument design.

Chromaplane is commercially produced and distributed by KOMA Elektronik.

Project Info
Artist: Passepartout Duo
Year: 2021 (development); 2024 (commercial launch)
Image Credit: Passepartout Duo / KOMA Elektronik

Friends. Lovers. Therapists. ‘Deathbots’. Artificial intelligence is now fulfilling new roles for millions of us every single day. How are these new ‘relationships’ changing how we view technology – and each other?

Beyond those who are using AI chatbots for administrative tasks, some people are now preparing to adopt children with their AI partners; others are reaching out to companies offering services to ‘resurrect’ deceased loved ones; others still look to bots to find treatment for their mental health issues.

In Love Machines, James Muldoon guides through these new forms of love, intimacy and connection, drawing on compelling interviews with users, developers and chatbots themselves. Along the way, he sheds light on the social conditions which have led to the exponential rise of the use of AI companions, and the unregulated corporations behind these technologies seeking to profit from users.

Inside Prague’s historic Old Town Market Hall, Signal Space has launched as a permanent venue dedicated to immersive and digital art. Its inaugural exhibition, ECHOES OF TOMORROW, assembles ten international artists and studios working across generative systems, laser environments, spatial audio, and data-driven visual media. On view through 31 March 2026, the exhibition unfolds as a sequence of large-scale installations embedded within the renovated hall.

This is not a collection of screens placed inside a historic shell. The works function as environments. Projection, light, sound, and computational processes operate at room scale. In several installations, generative logic and interactivity are not hidden layers—they determine how space behaves.

Systems, Not Screens
Computation sets the tone. Spanish studio Playmodes’ Signes runs on algorithmic processes that govern both image and sound. Sequences evolve according to coded rules rather than fixed timelines; each iteration is calculated, not replayed.

That systemic emphasis shifts form in Shohei Fujimoto’s intangible #form. Fine laser beams articulate geometric volumes in midair, appearing stable from one angle and dissolving from another. Light defines structure directly. Material and computation intersect in Crystal Pixels: Silent Reflections by the Preciosa Design Team. Digitally controlled lighting integrates with traditional Czech crystal, embedding programmable illumination within a craft lineage long associated with Bohemian glass.

Data becomes composition in Onformative’s Meandering River, which transforms datasets into evolving visual fields through algorithmic processing. Numerical input resolves as motion and pattern. Historical reference is reframed in Fighters / For Palestrina Sicut, a collaboration between Quayola and Max Cooper. Classical visual and musical structures are translated into digitally generated image and sound, processed through contemporary computational systems.
The pattern is consistent: projection is not display format; it is spatial logic. Sound is not accompaniment; it structures time.

Immersion as Infrastructure
The exhibition scales those systems to architecture. Czech digital artist Ondřej Zunka’s PHOTOSYSTEM II presents digitally constructed environments referencing natural systems through computational imagery. Synthetic landscapes unfold across the room rather than within a frame. In Weaving Nowhere, Belgian artists Porz A Park and Arnaud Gerniers combine spatial sound and visual structures into a multi-sensory field shaped by acoustic and optical composition. Interactivity sharpens that architectural logic in Body Sketches / Scale Studies by Zach Lieberman. Motion tracking converts visitor gestures into projected line drawings in real time. Movement alters output; the body enters the system.

Generative animation continues the emphasis on rule-based formation in Markos Kay’s aBiogenesis, where evolving patterns reference biological processes without direct simulation. At the largest scale, Nohlab’s EVERYTHING integrates projection and spatial sound into a continuous audiovisual environment occupying the room as a unified field. Projection systems, laser arrays, lighting rigs, and spatial audio operate as architectural components. The market hall’s masonry remains visible, framing computational media within a pre-digital civic structure. The contrast clarifies the premise: digital systems are not temporary overlays but spatial agents embedded in permanent infrastructure.

Exhibition as Platform
ECHOES OF TOMORROW foregrounds generative processes, laser technologies, spatial audio, and projection-based environments as primary artistic media. Algorithmic logic and interactivity are visible drivers of form rather than concealed mechanisms. As a permanent digital art venue, Signal Space provides the projection, lighting, and sound infrastructure required for immersive installations to function at architectural scale. The gallery is engineered for this mode of work.

The result is an exhibition organized around systems rather than objects. Projection defines volume. Laser defines boundary. Sound defines orientation. Computational processes shape how space is experienced.
Immersive media here operates as built environment. Not spectacle layered onto architecture, but infrastructure shaping it from within.

AI-Powered Conway’s Arcade is not a retro remake. Developed by SpecialGuestX and commissioned by Google, the project was presented at NeurIPS 2025 as an interactive installation that generates gameplay live rather than loading a fixed, preauthored title.

At first glance, the cabinet resembles a conventional upright arcade machine: joystick, mechanical buttons, compact footprint, pixel-style graphics. The difference lies in the software. Each session produces a distinct gameplay configuration in real time. Instead of replaying a static level, players encounter shifting arrangements of mechanics that echo classic arcade genres without replicating any single one. The project positions the arcade cabinet as a platform for exposing computational systems in action.

From Cellular Automata to Interactive Systems
The conceptual reference point is John Horton Conway’s Game of Life, the 1970 cellular automaton demonstrating how simple local rules can produce complex global patterns. In Conway’s Game of Life, cells switch states based on neighboring conditions. No central controller determines the outcome; structure emerges from rule interaction. AI-Powered Conway’s Arcade draws on this rule-based logic as inspiration. Rather than functioning as a conventional AI demo built around large-scale models or scripted scenarios, the cabinet generates gameplay configurations through structured rule interactions. Sessions unfold as dynamic arrangements of obstacles, agents, and motion patterns derived from this generative framework.

This approach distinguishes the project from standard procedural games. Traditional arcade software runs fixed binaries. Even procedurally generated titles typically operate within predefined templates and content libraries. Here, the emphasis is on recombining rule-driven behaviors at runtime, producing sessions that vary in structure rather than simply reshuffling assets. Players learn the system through interaction. Instead of following explicit instructions, they observe how elements behave within the current configuration. Patterns become legible through play.

Real-Time Variation, Legible Mechanics
Visually, the cabinet references early arcade graphics—flat color fields, geometric sprites, high contrast. The restrained aesthetic makes behavioral changes easier to register. When gameplay configurations differ between sessions, those differences are visible.

Individual runs may evoke familiar mechanics: paddle-and-ball exchanges, vertical obstacle navigation, projectile-based interactions. These are structural echoes rather than direct recreations. The system recombines recognizable arcade logics into new arrangements, producing experiences that feel referential without being fixed remakes. Because each session begins with a newly generated configuration, memorization alone offers limited advantage. Players instead attend to how movement, collision, and spatial constraints operate within the present setup. The cabinet foregrounds rule interaction as the primary site of engagement.

Hardware as Interface to Computation
The installation is housed in a custom-built aluminum cabinet designed for transport and exhibition contexts, including its presentation at NeurIPS 2025. While engineered for mobility, it retains the proportions and ergonomics of a traditional upright arcade machine.

The interface follows established conventions: joystick and mechanical buttons deliver immediate tactile response. The physical structure remains constant across sessions; what changes is the on-screen configuration. This separation between stable hardware and variable software underscores the project’s focus on generative systems. Installed within a research conference environment, the cabinet functioned as a playable demonstration rather than a commercial product. Instead of presenting computational models through diagrams or dashboards, the system was accessed through embodied interaction. Participants engaged with rule-based behavior directly, encountering generative logic as a playable system rather than an abstract concept.

AI-Powered Conway’s Arcade does not introduce a new arcade genre. It repositions the arcade machine as a container for real-time generative computation. By situating rule-based systems inside a familiar hardware typology, the project makes computational emergence observable through play—turning abstract logic into something tactile, visible, and testable.

name: Conway’s Arcade

agency: SpecialGuestx | @specialguestagram + @specialguestx
client: Google | @google
executive creative directors: Aaron A Duffy | @dufslam + Miguel Espada | @miguelvespada
creative directors: Marc Reisbig | @marc_reisbig + Carlos Font | @carlosfont
executive creative technologist: Federico Guardabrazo | @guardabrazo
user experience: Laura Fajardo | @laurafajardoc
lead producer: Francesca Palau del Mas | @francesca_pdm
lead creative coder: Bruno Barrán | @brunobarran
executive visual identity direction: 1stAve.Design | @1stave.design
visual identity: Polar, Ltda. | @polar.ltda
lead product designer: Seis Punyales | @seispunyales
product producer: Alicia Simon | @aliciasimon
product designer: Mike Fernández | @mikes.ink
product designer: Alicia Simon | @aliciasimon
product designer: Paco Fuster | @fuster.ferrer
lead electronics engineer: Alvaro Gordo | @creativengineer.jpg
technical consulting: Pedro Zambrana | @pedrozambrana
graphic design: Eder Larrondo | @edd.er
fabricated in barcelona by: Seis Punyales | @seispunyales
photographer: Paula Vázquez Guisande | @paulavguisande

At the February 2026 edition of New York Fashion Week, designer Maia Hirsch presented a garment that does more than animate—it responds. Her “Blooming Dress” (Peace Offer) mechanically opens when the wearer shakes someone’s hand. Embedded touch sensors detect physical contact, triggering internal actuators that expand petal-like structures outward from the body. In the same presentation, a second piece illuminated in response to motion, using integrated lighting activated by movement.

The engineering is compact and deliberate. Sensors, control circuitry, micro-actuation systems, and onboard power are integrated within the garment’s structure, enabling repeatable performance under runway conditions. Rather than treating electronics as decorative add-ons, the pieces operate as cohesive wearable systems: sensing input, processing signals, and executing mechanical output in real time.

Interactive fashion has precedents. Hussein Chalayan famously presented motorized dresses that transformed shape on the runway in the mid-2000s. Iris van Herpen is well know for having collaborated with engineers and architects to develop structurally dynamic couture that pushes material fabrication into computational territory. What differentiates Hirsch’s NYFW presentation is its use of social interaction as the trigger condition. A handshake becomes a switch. Proximity becomes input.

Does the high-profile runway, in this context, function as spectacle or more as a controlled deployment environment for experimental hardware? Garments are increasingly assembled through hybrid workflows that merge pattern-making software with PCB layout, conductive thread routing, and parametric modeling. The result is clothing that behaves less like static fabric and more like distributed interface. Wearable robotics has often been framed as futuristic theater. At NYFW 2026, it appeared as systems design operating at body scale.

The phrase “post-digital craft” is often misread as retreat—a soft-focus return to hand tools and pre-industrial methods. In practice, it signals something more structural: a recalibration of making after digital tools became default infrastructure. Post-digital craft does not reject computation. It absorbs it, retools it, and insists that technique, material intelligence, and embodied judgment still matter—even when the tools are robotic arms and parametric scripts.

As AI-generated imagery saturates feeds and manufacturing pipelines grow increasingly automated, a countercurrent is visible across design studios and fabrication labs. Authors are not abandoning machines. They are relocating authorship into systems—into parameters, toolpaths, calibration routines, and fabrication protocols. The result is hybrid practice: neither analog revival nor techno-optimism, but materially grounded and computationally fluent.

Hybrid Making After Automation
Digital fabrication expanded rapidly in the 2010s. Fab labs and makerspaces standardized access to CNC routers, laser cutters, and desktop 3D printers. That expansion democratized production—but it also normalized a recognizable visual language: parametric surfaces, algorithmic lattices, and toolpath aesthetics. Post-digital craft begins after that normalization. The machine is no longer the spectacle. The focus shifts to tuning—surface behavior, movement, structural performance—through iterative testing rather than frictionless output.

Amsterdam-based Studio Drift exemplifies this shift. Founded by Lonneke Gordijn and Ralph Nauta, the studio builds kinetic installations that combine custom engineering with material experimentation. Their work Shylight unfolds and retracts in sequences inspired by nyctinasty, the movement of flowers. The installation relies on precisely calibrated mechanical systems, yet the perceptual effect emphasizes timing, sensitivity, and variation. Technology functions as a tuned instrument rather than a productivity engine. The emphasis is not on speed or scale. It is on calibration.

Craft Knowledge in Code
If traditional craft has been described as tacit knowledge accumulated through repeated manual practice, contemporary digital fabrication distributes that expertise across software environments and robotic systems. In computational workflows, decisions about geometry, density, and structural reinforcement are encoded before material is shaped. Skill does not disappear; it migrates upstream.

Dutch designer Joris Laarman has explored this migration for more than a decade. His collaboration with MX3D on the MX3D Bridge in Amsterdam used robotic wire-and-arc additive manufacturing (WAAM), a process that deposits molten metal through an electric arc. WAAM requires precise control of heat input, deposition rate, and structural stability. The bridge’s form emerged through generative modeling and structural analysis, then underwent empirical testing to ensure performance. Here, the designer configures systems rather than directly manipulating material. Robotic arms execute programmed instructions, but human operators define parameters, oversee calibration, and evaluate outcomes. Iteration remains central: simulation informs fabrication; fabrication feeds back into code.

This redistribution of expertise alters how error functions. In software-driven workflows, modeling inaccuracies can propagate through fabrication if left unchecked. As a result, prototyping, mock-ups, and sensor data remain essential. Digital precision does not eliminate craft judgment. It makes it procedural. Institutions have acknowledged this shift. The Victoria and Albert Museum (V&A) has collected and exhibited works incorporating generative design and digitally fabricated components, including projects by Laarman. By situating these works within decorative arts and design collections, the museum frames robotic fabrication as part of a longer history of material practice rather than as a break from it.

Why Post-Digital ≠ Anti-Digital
The “post” in post-digital does not mean after technology. Media theorists have used the term to describe a condition in which digital systems are normalized—embedded so deeply in production and culture that they become infrastructural. The question is no longer whether to use computational tools, but how they are configured and governed. Post-digital craft rejects two simplifications: the assumption that automation inherently produces better outcomes, and the idea that authenticity lies only in manual labor. Instead, it treats computation as material—something to shape, constrain, and interrogate.

This has implications beyond aesthetics. Research in distributed manufacturing and circular design suggests that localized, small-batch production—when combined with repairability and modularity—can reduce transportation and tooling waste under certain conditions. Hybrid workshops that combine CNC machining with hand-finishing or robotic printing with manual assembly introduce adaptability into systems often optimized for scale. These outcomes are contingent, not guaranteed. But they demonstrate that digital fabrication is not structurally tied to mass standardization.

Time also re-enters the equation. Sociological analyses of technological acceleration describe how modern systems compress production cycles. Yet hybrid digital–manual workflows still require machine calibration, simulation testing, and physical prototyping. These phases slow processes down—not as inefficiency, but as method. Iteration becomes a site of evaluation rather than delay.The rise of generative AI intensifies this terrain. As automated systems produce images, text, and product concepts at scale, debates around authorship and agency expand. Post-digital craft responds by embedding authorship within process design—within how systems are structured, constrained, and executed. Originality shifts from surface novelty to procedural specificity.

Post-digital craft is not a revival of pre-digital romanticism. It is a condition in which computation is fully integrated into materially grounded practice. Silicon and steel coexist. Script and surface inform one another. Machines do not replace judgment; they expose where judgment resides. The nostalgia narrative misses the point. What is emerging is not a retreat from technology, but a demand that technology remain accountable to material intelligence, human calibration, and sustained attention.

Scientists in Bristol, England have demonstrated a prototype shape-shifting robot that deforms and moves under the influence of external electric fields, marking a noteworthy advance in soft robotics and electro-responsive materials research. The robot is constructed from a novel electro-morphing gel that reacts to changes in electric field configuration around it, rather than relying on conventional motors or rigid mechanical parts.

The prototype — visually likened to a “gymnast” during demonstrations — bends, stretches, twists and swings when electrodes around its body are energized in specific patterns. By altering the spatial distribution and intensity of these fields, researchers can prompt the material to undergo controlled deformations that translate into locomotion and shape change. This approach exploits the gel’s inherent material properties, which respond directly to electric forces rather than through intermediary mechanical linkages.

The development originates from work at the University of Bristol School of Engineering Mathematics and Technology, where engineers have been exploring how electro-responsive composites can be programmed to generate complex motion patterns. According to the report, the robot’s design leverages external electric field control to manipulate its morphology, rather than embedding actuators or circuitry within the body. This external actuation method simplifies the robot’s physical architecture but also highlights the challenge of integrating control systems for autonomous operation.

Shape-shifting robots like this prototype align with broader trends in soft robotics, where compliant structures and responsive materials replace traditional joints and rigid frames to enhance adaptability when interacting with unstructured environments. While practical applications remain exploratory, potential use cases include systems that must traverse confined or irregular spaces or interfaces that adapt dynamically to tasks without complex mechanical control layers.

The Bristol work underscores how material-centric design, paired with externally programmable fields, can expand how researchers think about robotic actuation and form modulation, particularly where conventional robotics faces limitations.

For years, computational design has been defined by modeling: parametric scripts generating geometry, simulations optimizing performance, machine learning systems producing form variations. Matter entered at the end of the pipeline—fabricated after computation had already concluded. Physical computation reverses that sequence. When behavior is encoded directly into material structure, topology, and compliance, computation no longer sits exclusively in silicon. It unfolds through deformation, phase change, distributed actuation, and multi-stable geometry. Under these conditions, design practice shifts. If materials compute, designers are no longer producing static forms—they are authoring behavioral systems.

Encoding Behavior in Matter
At the forefront of this shift is Lining Yao and her Morphing Matter Lab at Carnegie Mellon University. Yao’s research focuses on materials that change shape in response to environmental stimuli such as water and heat. Her team has developed 3D-printed hydrogels and bio-based composites whose microstructures determine how they swell, curl, or fold when exposed to moisture. In these systems, transformation is not animated externally. It is embedded in composition and geometry. Print patterns, fiber orientation, and material gradients encode how the structure will respond. Once fabricated, the object carries its own conditional logic.

This approach alters the workflow fundamentally. Instead of modeling a final geometry and simulating its performance, the designer defines a transformation rule: how much swelling occurs, along which axis, at what threshold. Form becomes a temporal expression of material instruction. Fabrication is no longer downstream of computation; it is computation. Such systems also operate with minimal external energy. Hygroscopic and hydrogel-based responses rely on environmental input rather than continuous electrical power. The substrate encodes its response behavior in material structure.

Projects such as MorphingSkin extend the lab’s morphing-material research into thin, programmable surface systems, where patterned structures encode how a material curls, folds, or textures in response to environmental input.

Topology as Instruction Set
While morphing materials encode behavior through chemistry and microstructure, programmable mechanical metamaterials encode behavior through geometry. At ETH Zurich, Kristina Shea leads research into multi-stable structures whose topology determines how they move and reconfigure under load. These latticed systems can snap between stable states without continuous energy input. In many cases, their behavior is governed by the arrangement of beams, hinges, and cellular patterns rather than continuous motorized control. A shift in topology alters how stress propagates through the structure. Multi-stability is achieved through structural design rather than embedded electronics.

In this model, geometry determines the system’s possible state transitions. The designer specifies relationships—angles, thicknesses, and connection points—that define how the structure responds to applied force. Once fabricated, the system behaves according to its geometry and material properties; its response is inseparable from its structural logic. For practice, this demands a shift from composing surfaces to engineering possibility space. The designer is not asking what shape to produce, but what behaviors to enable. In multi-stable metamaterials, analysis and fabrication converge: topology carries the behavioral rule.

The implications extend beyond laboratory prototypes. Such structures suggest architectural components, deployable systems, and adaptive assemblies that respond passively to load and environment. Maintenance, durability, and tolerance stacking become central design concerns. Wear and fatigue are operational variables embedded within the system’s long-term behavior.

Distributed Intelligence in Soft Robotics
If Yao’s work embeds behavioral response in material microstructure and Shea’s encodes it in topology, the soft robotics research of Daniela Rus shows how embodied mechanics scale into autonomous systems. At MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL), Rus and collaborators have developed both soft robotic systems—such as pneumatically actuated grippers—and self-reconfigurable modular robots like the M-Blocks platform. These projects explore how structure, material compliance, and distributed interaction contribute to adaptive behavior.

In soft robotics, body and control are tightly coupled: morphology and material dynamics can simplify control requirements by shaping how forces propagate and how the system responds to contact. Feedback emerges from interaction between material and environment, enabling behaviors that are difficult to achieve with rigid mechanisms alone. Designing these systems means configuring how sensing, actuation, and structure interrelate. Creative control shifts from sculpting a static object to specifying behavioral coupling—compliance gradients, actuator placement, and modular connectivity—then iterating through physical calibration as the robot’s behavior is validated in real operating conditions.

From Artifact to System
Across these examples, a consistent pattern emerges. The designer’s task is no longer limited to producing a finished artifact. Instead, it involves defining a rule-bound system that continues to operate after fabrication. This transformation has practical consequences. Iteration cycles expand beyond digital simulation into material calibration. Environmental variability—humidity, load, temperature—affects output. Tolerances, fatigue, and degradation become computational parameters. The system’s lifespan is part of its logic.

Physical computation also complicates narratives of decentralization. Many responsive systems still rely on embedded electronics and complex supply chains. Multi-stable structures may eliminate continuous energy input but require precision manufacturing. Soft robots reduce control overhead yet depend on sophisticated materials research. The shift from form-making to rule-making does not remove control; it redistributes it across structure, material, and feedback loops.

At the same time, as artificial intelligence increasingly automates symbolic production—generating images, text, and parametric scripts—physical computation anchors intelligence in tangible substrates. It positions material systems as active participants in sensing and response. In this framework, the designer becomes less a stylist of surfaces and more an author of conditions. Material choice is a computational decision. Topology is a behavioral script. Fabrication is an encoding process. When matter computes, design becomes the practice of shaping how systems act—not just how they look.

Researchers at The University of Texas at Austin, working with Sandia National Laboratories and other U.S. laboratories, have developed a 3D printing technique that enables fine-grained control over the internal structure of thermoplastics. The method, called Crystallinity Regulation in Additive Fabrication of Thermoplastics (CRAFT), allows printed objects to vary in stiffness, flexibility, and optical properties within a single material system.

CRAFT uses digital light processing (DLP) 3D printing and a widely available liquid resin, cyclooctene, that can be polymerized and crystallized in a controlled manner. By modulating grayscale light patterns during printing, the researchers adjust the degree of crystallinity at specific locations inside the object. Because crystallinity directly influences mechanical performance, this approach enables spatial tuning of properties without switching materials or print heads.

In demonstrations reported by the university, the team fabricated a model human hand with regionally differentiated mechanical behavior, approximating the relative rigidity of bone and the flexibility of surrounding tissues such as ligaments and tendons. A grayscale design controlled light intensity during printing, producing distinct crystalline and amorphous domains within a single feedstock material. Traditional multi-material 3D printing can achieve visual complexity by combining different inks, but these materials often adhere poorly at their interfaces, leading to structural weaknesses. The CRAFT method instead varies properties by regulating crystallinity during fabrication, eliminating the need to bond dissimilar materials and reducing interfacial failure.

The technique is compatible with commercially available digital light processing (DLP) or LCD printers capable of grayscale projection, including relatively low-cost systems. According to the university, potential applications include anatomical training models for medical education and energy-damping structures such as personal protective gear. The researchers also point to bioinspired materials with alternating hard and soft regions, similar to structures found in bone or tree bark, as another area of exploration. By enabling spatial control over mechanical and optical properties within a single printed object, CRAFT extends additive manufacturing beyond geometric complexity toward internal structural differentiation.

Cognitive Bloom is a speculative design concept developed through a collaboration between Chanwoo Lee, Map Project Office, and Lovelace Research. Rather than positioning artificial intelligence as a productivity accelerator, the project proposes an alternative: AI as a quiet companion for structured self-reflection.

The concept centers on an ambient display that visualizes personal wellbeing data as a living digital ecosystem. Areas of strain appear as yellowing leaves. Signs of progress surface as new buds. When aspects of wellbeing stabilize, those buds bloom. The interface avoids dashboards, scores, or clinical-style metrics. Instead, it translates abstract data into a metaphor most users intuitively understand: tending a garden.

Cognitive Bloom is presented as a domestic object designed to encourage what its creators describe as a ritual of reflection. Interaction is deliberately slowed. The system resists the logic of instant answers and frictionless optimization common to many contemporary AI tools. Rather than compressing cognition into efficiency metrics, it introduces pauses—moments intended for observation and consideration. The gardening metaphor extends throughout the project’s framing. Gardens demand attention over time. They respond to changing conditions. They cannot be rushed. In this way, Cognitive Bloom contrasts with engagement-driven digital systems that prioritize speed, scale, and behavioral capture.

As a speculative project, Cognitive Bloom does not claim full technical implementation. Instead, it operates as a design probe, asking how artificial intelligence might be shaped around wellbeing rather than extraction. By situating AI within a metaphor of care and cultivation, the project reframes the role of intelligent systems in domestic life—not as engines of acceleration, but as structures for presence.

Cultural circulation online is often framed as frictionless distribution: publish, post, upload, and consumed. In practice, visibility is governed by ranking systems, recommendation engines, and moderation pipelines that sit between creators and publics. Platforms such as TikTok, Instagram, and YouTube operate less like neutral conduits and more like infrastructural curators, determining what surfaces, what trends, and what disappears.

Algorithmic visibility is not simply a technical feature. It is a system of cultural governance embedded in code, policy, and economic incentives. Understanding who controls cultural circulation online requires examining how these systems rank, filter, monetize, and increasingly generate content.

From Chronological Feeds to Algorithmic Governance
Many major social platforms originally defaulted to reverse-chronological feeds, but over time shifted toward recommendation-driven ranking optimized for engagement signals. YouTube has described its recommendation system as using signals including clicks, watch time, and measures of viewer satisfaction derived from surveys. Reports have attributed a widely cited figure—around 70% of watch time—to recommendations, based on public remarks by YouTube’s leadership at industry events.  TikTok’s “For You” feed similarly ranks videos using factors such as user interactions, strong interest signals like watch completion, video information (captions/sounds/hashtags), and some device/account settings. Instagram, once anchored in a follower-based chronological feed, shifted to algorithmic ranking and has been described (including by platform explainers and reporting around the 2016 change) as prioritizing predicted interest, relationship, and timeliness when ordering content.  These systems do not merely sort content; they shape the conditions under which posts become discoverable.

Optimization as Aesthetic Constraint
Because recommendation systems prioritize metrics such as watch time, retention, and engagement, creators often adapt form and pacing to align with those signals. YouTube has publicly identified watch time and viewer satisfaction as central to its recommendation system. TikTok’s For You feed emphasizes watch completion and interaction signals, while Instagram has described how Reels and other recommendation surfaces rank content based on predicted interest and engagement.

Creator guidance materials frequently emphasize click-through rates, retention curves, and early engagement signals, encouraging thumbnails, titles, and opening seconds that maximize viewer response. When ranking systems reward frequent posting, high retention, and specific formats, creators adjust accordingly. Researchers and commentators have observed that this dynamic can contribute to format convergence over time, as visibility becomes tied to measurable engagement patterns. Monetization systems reinforce these incentives. YouTube’s Partner Program requires minimum subscriber and watch-hour thresholds before revenue sharing is enabled, creating economic gates that influence growth strategies. Visibility and monetization are structurally intertwined: ranking affects revenue potential, and revenue potential shapes production decisions.

The Politics of Amplification and Omission
Algorithms amplify some material while marginalizing other forms. This dynamic extends beyond virality to include classification, moderation, and training data. Artist and researcher Mimi Ọnụọha has examined how datasets reflect structural absence, arguing that what is missing can be as consequential as what is included. In the context of platform algorithms, omissions may stem from biased training data, automated moderation errors, or ranking systems whose optimization criteria deprioritize certain types of content. Platforms deploy automated moderation tools to detect hate speech, misinformation, and other policy violations, and publish transparency reports outlining enforcement metrics. However, the specific functioning of these models remains largely proprietary. Independent research has documented false positives and uneven enforcement, including cases in which language associated with marginalized communities is misclassified by automated systems.

Amplification also intersects with generative AI systems. Large language models and image generators are trained on vast corpora of publicly available data. Because widely circulated and highly indexed content is more likely to be scraped into training datasets, recommendation systems can indirectly influence the composition of future training data. This creates a recursive dynamic in which ranking systems shape what becomes broadly available cultural material, which in turn informs subsequent generative outputs. Such feedback loops can consolidate influence within infrastructures that are neither neutral nor evenly distributed.

Artists and Designers Working on the System
Some practitioners have treated algorithmic systems themselves as material. Hito Steyerl has examined how images circulate through compression, duplication, and digital distribution networks, tracing the relationship between visibility and power. Her essay “In Defense of the Poor Image” analyzed how distribution conditions shape meaning and authority. James Bridle has investigated opaque technological systems, including automated decision-making and machine vision, emphasizing the difficulty of understanding infrastructures that shape public life. Through research and artistic intervention, his work renders such systems visible and frames algorithms as active participants in cultural production rather than neutral utilities.

Artists including Tega Brain and Sam Lavigne have engaged directly with APIs, data scraping, and automated platforms, exposing how networked systems sort, classify, and monetize attention. Their projects treat code, metadata, and platform rules as sites of intervention, demonstrating that circulation itself is programmable. These practices do not dismantle platform power, but they surface its mechanics. By foregrounding the infrastructures behind recommendation engines, moderation systems, and data economies, they highlight the importance of understanding how algorithmic systems shape contemporary cultural circulation.

Who Controls Cultural Circulation?
Control over cultural circulation online is shaped by multiple actors, yet concentrated within a small number of dominant platforms. Platform companies design and continually adjust ranking systems. Advertiser requirements influence monetization policies and brand-safety standards. Machine learning engineers tune models using engagement and safety metrics. Users generate the behavioral data that enables personalization. Generative AI systems are trained on large corpora of publicly available content that reflect existing patterns of visibility and circulation.

The result is a multi-layered governance structure in which visibility emerges from interactions between code, policy, and market incentives. Cultural production is shaped at multiple points by recommendation systems and monetization thresholds. Algorithmic visibility has become a structural feature of contemporary networked platforms. As recommendation engines, automated moderation systems, and generative models become more deeply integrated, the central question is not simply whether algorithms influence circulation, but how that influence is designed, audited, and contested.

Georges’ broader practice spans literature, sound, photography, video, and 3D modeling. Since her debut novel La Mue de l’hermaphrodite (2001), she has examined how emerging technologies reshape concepts of subjectivity, memory, and embodiment. In recent years, her work has increasingly engaged artificial intelligence not as a novelty, but as a perceptual framework.

In Woven Presence, surface operates as both material reference and computational field. Textile logic—repeat, weave, ornament—intersects with generative image systems that mutate texture and form over time. Figures appear embedded within, or produced by, their own patterned environments. The distinction between fabric, skin, and simulation becomes visually unstable.

The project foregrounds mediation. Machine vision reorganizes the sensible world: tactility is rendered as pixel density; ornament becomes data structure; portraiture shifts toward probabilistic construction. The result is a work that situates intelligent systems within longer histories of image-making and material culture.

Woven Presence contributes to ongoing conversations in digital art and creative technology about authorship, hybridity, and the evolving relationship between algorithmic systems and embodied perception.

Project Info
Artist: Karoline Georges
Year: 2023
Image Credit: Courtesy of the artist

Design culture has spent decades optimizing for novelty: faster tools, smarter systems, more autonomous outcomes. But as infrastructures age and environmental volatility intensifies, the limits of mastery-first design are increasingly visible. Buildings, platforms, and material systems rarely fail because they were insufficiently clever; they fail because they were insufficiently maintainable. Designing for maintenance reframes durability not as permanence, but as the capacity to endure breakdown, repair, and care over time.

Across architecture, materials research, and community-led practice, a quieter design ethic is emerging—one that treats upkeep as a primary condition rather than a postscript. This approach resists speculative solutionism and instead foregrounds stewardship, repair cultures, and systems designed to fail gracefully.

Maintenance as a Primary Design Condition
Maintenance has long been treated as an operational concern, external to the design act itself. Yet research-driven practices are increasingly exposing how maintenance is inseparable from form, material choice, and system logic. Architect and researcher Lydia Kallipoliti has been central to this reframing. Through projects such as Closed World and her broader scholarship on material ecologies, Kallipoliti examines closed-loop systems not as seamless technological fixes, but as inherently leaky, contingent constructs. Her work demonstrates that circular material systems demand constant intervention: monitoring material fatigue, recalibrating flows, and managing inevitable loss.

This perspective matters because it destabilizes the myth of self-sustaining systems. Closed loops do not eliminate maintenance; they intensify it. Designing with recycled or hybrid materials introduces variability that must be anticipated, not engineered away. For designers, this shifts the goal from perfect control to informed custodianship—understanding how materials age, deform, and require care. Maintenance, in this sense, becomes a design parameter that shapes spatial logic and material expression. Systems are legible rather than hidden. Interfaces expose wear rather than conceal it. The result is architecture and infrastructure that communicates its own vulnerability—and, crucially, how it can be sustained.

Long-Term Stewardship vs. Speculative Solutionism
Speculative design has played a valuable role in expanding the imagination of what systems could be. But when speculation hardens into solutionism, it often prioritizes dramatic intervention over lived continuity. Community-centered practices offer a counterpoint. The UK-based collective Assemble is frequently cited for its emphasis on process over product. Projects such as the Granby Four Streets regeneration in Liverpool were not conceived as finished solutions but as frameworks for ongoing repair. Assemble’s role was not to replace community knowledge with professional authority, but to scaffold conditions under which residents could continue to adapt, fix, and maintain their environment.

What distinguishes this approach is its temporal logic. Success is not measured at completion but years later, through occupancy, modification, and care. Materials are chosen for accessibility rather than novelty. Construction methods favor techniques that can be relearned and repeated by non-specialists. Maintenance is social as much as technical. This model exposes a blind spot in much speculative work: systems designed without clear stewards tend to externalize responsibility. When maintenance is someone else’s problem, failure becomes more likely. Assemble’s work demonstrates that durability emerges from shared ownership and embedded repair cultures, not from optimization alone.

Designing Systems That Endure Breakdown and Repair
If maintenance is inevitable, then breakdown must be treated as a design scenario rather than an exception. Few architectural programs embody this logic as consistently as Rural Studio, the long-running design-build initiative based in Alabama. Operating in contexts of limited resources, Rural Studio designs buildings that anticipate wear, modification, and local repair. Their projects are intentionally over-documented and materially straightforward, enabling residents to maintain structures long after the designers have left. The architecture does not assume stable funding streams or specialized maintenance crews; it assumes improvisation.

This orientation toward continuity through care produces a different aesthetic and technical outcome. Details are robust rather than delicate. Systems are redundant rather than optimized to the edge. Materials weather visibly, making deterioration legible rather than catastrophic. In doing so, Rural Studio challenges the notion that resilience requires advanced technology. Often, it requires humility about what can realistically be sustained. For those working in digital or hybrid systems, the lesson translates directly. Platforms and tools that assume uninterrupted uptime or ideal user behavior tend to fracture under real-world conditions. Designing for maintenance means accounting for partial failure, degraded performance, and user-driven repair—whether that repair is technical, social, or organizational.

From Mastery to Care
Designing for maintenance is not anti-innovation; it is anti-amnesia. It recognizes that every system persists beyond its launch moment and that endurance is shaped less by brilliance than by care. Closed-loop materials, community-led repair, and design-build stewardship all point toward a shared conclusion: continuity is produced through ongoing attention.

As climate instability, resource constraints, and infrastructural fragility become baseline conditions, the most relevant design intelligence may lie not in predictive mastery but in adaptive maintenance. Systems that acknowledge their own fragility—while making repair possible—are better equipped to survive the long arc of use. In this framing, maintenance is not a compromise. It is a form of authorship that unfolds over time, shared among designers, users, and caretakers alike.

Researchers at the University of Hong Kong have developed what they describe as the world’s first soft, three-dimensional, biocompatible semiconductor made from hydrogel, marking a significant step toward electronics that physically and functionally integrate with living tissue. The work was led by Shiming Zhang and his team in HKU’s Wearable, Intelligent, Soft Electronics (WISE) research group.

Unlike conventional semiconductors, which are rigid, flat, and mechanically mismatched with the body, the HKU device is water-rich, deformable, and volumetric. The hydrogel-based semiconductor can be fabricated into thick, three-dimensional structures while remaining soft—closer in texture and mechanical behavior to biological tissue than to silicon. According to the researchers, the material is also compatible with living cells, which can survive and grow within the electronic structure itself.

The team demonstrated three-dimensional hydrogel transistors capable of conducting electrical signals while hosting living cells, a combination that has not been achievable with previous electronic materials. This enables electronic architectures that operate not only on the surface of tissue but within tissue-like volumes, opening new possibilities for bioelectronic interfaces. The research, reported in Science, positions hydrogel semiconductors as a potential foundation for future neural interfaces, prosthetics, and regenerative medical technologies. Because the material reduces mechanical mismatch and supports biological integration, it addresses longstanding challenges in implantable electronics, including inflammation, signal degradation, and device longevity.

Rather than presenting electronics as external tools applied to the body, the HKU work reframes electronic systems as materials that can coexist with biological structures. As interest grows in brain–machine interfaces, soft prosthetics, and biointegrated healing technologies, this development points toward a class of electronics designed to blend into living systems rather than resist them.

At MIT, researcher and designer Cyrus Clarke has developed Anemoia, an experimental device that explores whether visual images can be translated into olfactory outputs through computationally mediated systems. Situated at the intersection of design research, sensory interfaces, and machine learning, Anemoia is not a commercial product or a claim about automating fragrance creation. It functions as a research prototype that examines how smell—one of the least standardized human senses—might be structured and mediated through technical systems.

Anemoia takes photographs as input and produces corresponding scent outputs via a controlled emission mechanism. Visual features are computationally mapped to olfactory representations informed by learned associations between images, descriptive language, and scent references. Rather than generating a single definitive smell, the system produces an approximation shaped by probabilistic translation rather than direct sensory measurement.

Smell as a Computational Problem
Olfaction has historically resisted digitization. Unlike vision or sound, it lacks a shared representational framework. There is no universal scale for smell, and language used to describe it varies widely across cultures and individuals. These inconsistencies make smell difficult to encode, compare, or reproduce computationally.

Anemoia does not attempt to resolve these ambiguities. Instead, Clarke’s work treats them as constraints. The system operates on the premise that smell can be approximated indirectly, through correlations rather than direct measurement. Visual cues—such as color, texture, and object category—are computationally associated with olfactory descriptors drawn from human-labeled references, which inform the system’s scent outputs. In this structure, smell is treated as probabilistic rather than absolute. A photograph of a forest path may align with descriptors such as woody or earthy, which map to certain aroma molecules. The system does not claim that this is how the scene “actually” smells. It reflects how such scenes are commonly described within the data it draws from.

How Anemoia Works
Technically, Anemoia combines image analysis with computational mapping to olfactory components. Images are processed to extract visual features, which are then associated with scent descriptors through trained models. These descriptors inform the system’s scent outputs, which are rendered through a physical emission mechanism. The scent output is delivered via a custom-built hardware interface designed to release controlled quantities of fragrance materials. This physical component is essential to the project. Anemoia is not purely a software model; it is a device that closes the loop between computation and embodied experience. The emitted scents allow users to encounter the system’s translations directly, rather than as abstract data.

Clarke has framed Anemoia as a design probe rather than a finished system. The device exposes the assumptions embedded in its mappings: which visual cues are prioritized, which scent descriptors are available, and how cultural conventions shape the dataset. These choices influence the resulting smells, making the system’s biases perceptible rather than hidden.

Designing Smell as an Interface
By translating images into olfactory outputs, Anemoia reframes smell as an interface layer. Scent becomes something that can be triggered, modulated, and synchronized with other media, rather than remaining ambient or incidental. This framing aligns olfaction with interaction design, where sensory outputs are treated as part of a broader system rather than isolated effects.

The project also foregrounds authorship at an infrastructural level. Decisions about datasets, descriptors, and mappings shape how the device interprets images and what it produces in response. These decisions are not neutral. They embed particular cultural and sensory norms into the system, determining which smells are legible and which are excluded. Anemoia does not suggest that machines can capture the fullness of human smell perception. Instead, it makes visible the gap between lived experience and computational representation. By doing so, it offers a way to examine how emerging technologies may increasingly mediate sensory experience—not by replicating it, but by approximating it through systems of translation.

As design and technology move further into multisensory territory, projects like Anemoia highlight the importance of treating perception itself as a designed system. Smell, once considered too subjective to model, becomes a site where technical, cultural, and aesthetic decisions converge.

Tezos and HEK Basel have announced a digital art partnership that brings blockchain into an institutional, research-oriented context within contemporary art. The collaboration connects Tezos, a proof-of-stake blockchain, with HEK Basel, a leading European institution dedicated to new media art, digital culture, and experimental practice.

HEK Basel has a long history of exhibiting and contextualizing work that engages software, networks, and emerging technologies within broader cultural and political frameworks. In this context, the partnership reflects a broader shift toward institutional and research-led engagement with blockchain, positioning it as part of digital art practice rather than primarily as a speculative market mechanism. Under the collaboration, Tezos will support HEK Basel’s institutional programming related to digital and blockchain-based art. Public materials emphasize engagement with decentralized technologies in artistic and curatorial contexts, alongside knowledge exchange and public-facing initiatives that address blockchain’s cultural implications.

For Tezos, the partnership aligns with a strategy of engaging established cultural institutions as part of its broader arts ecosystem. For HEK Basel, it provides resources to explore blockchain within existing curatorial and research frameworks while maintaining institutional independence. More broadly, the announcement points to a slower, institutionally grounded pathway for blockchain in the arts—one centered on sustainability, critical inquiry, and long-term cultural relevance. For artists and designers, it suggests a future in which decentralized technologies are treated as cultural infrastructure rather than short-term trends.

The installation originated within the eponymous music-theatre production directed by Ruben van Leer, with music by Tigran Hamasyan, which premiered at Holland Festival on June 8, 2024. Within that performance context, sculpture, text, sound, and choreography formed an integrated narrative environment.

The project references mythological traditions surrounding the firebird—figures associated across cultures with renewal and awakening. In Armenian tradition, the bird’s power resides in its voice. The installation extends this mythology into mechanical form: articulated wing structures suggest Leonardo da Vinci’s speculative flying machines, foregrounding the persistent human ambition to engineer flight.

Materially, the work stages a dialogue between natural and artificial systems. Fabric surfaces behave like membranes; motorized components impose programmed movement; tungsten light activates depth and shadow within ecclesiastical architecture, as seen in its 2026 presentation at Vilnius Light Festival in St. Catherine Church.

Rather than resolving illusion and engineering, the installation sustains their tension. The bird is neither purely symbolic nor purely technical—it is a constructed apparatus that makes visible the mechanics behind transcendence.

Project Info
Artist: Boris Acket
Year: 2024
Image Credit: Courtesy of the artist / production team (Truth.io)
Creative Direction: Boris Acket, Ruben van Leer
Inspired by the music of: Tigran Hamasyan
Creative Engineer: Merijn Versnel
Production: Truth.io
Supported by: Mondriaan Fund
Shown at: Vilnius Light Festival (2026), St. Catherine Church
Creative Production: Luiza Guidi
Software: Moos Rebolder

The Island is a site-specific project by Hito Steyerl, presented at Osservatorio, the exhibition space overlooking Milan’s Galleria Vittorio Emanuele II. Developed in close dialogue with the architecture and symbolic charge of its location, the work extends Steyerl’s sustained inquiry into how power operates through images, platforms, and infrastructures. Rather than isolating content from context, The Island treats the exhibition site as an active system—one shaped by visibility, circulation, and economic control.

The project can be read as mobilizing the island not as a place of retreat or autonomy, but as an operational condition. Across contemporary geopolitics, finance, and technology, islands function as zones of exception: special economic areas, offshore data centers, gated territories, and logistical enclaves. Steyerl’s installation foregrounds questions around how these spaces are produced, who benefits from their separation, and how their apparent isolation masks deep integration with global systems.

Osservatorio as an Active Infrastructure
Osservatorio’s position above the Galleria is central to the work. The space is elevated, transparent, and carefully controlled—neither fully public nor conventionally institutional. Below, the Galleria operates as a dense node of luxury retail, tourism, and urban spectacle. Rather than attempting to neutralize this setting, The Island can be read as leveraging it to foreground how architecture structures perception. The installation situates viewers in a position of observation that echoes contemporary systems of oversight: platforms that promise openness while managing access, attention, and value behind the scenes. Glass walls, sightlines, and spatial thresholds function as material conditions through which the work is encountered, situating visitors within a layered environment of visibility and exclusion.

This approach aligns with Steyerl’s broader practice, in which exhibition spaces are frequently treated as operative environments rather than neutral containers. At Osservatorio, the building’s hybridity—a cultural venue embedded within a commercial landmark—offers a case through which to consider how institutions participate in the systems they host.

The Island as a Logic of Separation
Rather than narrating a single story, The Island assembles fragments that can be read as pointing to separation as an infrastructural strategy. Visual and spatial elements emphasize enclosure, segmentation, and controlled access. The island emerges here as a logic through which contemporary power operates: carving out zones that appear autonomous while remaining tightly coupled to global flows of capital, data, and labor. Steyerl’s avoidance of linear explanation is consistent with her broader practice. The work does not present dashboards, speculative interfaces, or invitations to optimize or manage systems. Instead, it positions the viewer as an interpreter confronted with partial views and mediated perspectives—an experience that echoes how many real-world infrastructures function: opaque, distributed, and resistant to singular points of control. In this sense, The Island can be understood less as representation than as reenactment. Moving through the installation parallels forms of contemporary governance in which decisions are registered spatially and affectively, even as their mechanisms remain out of reach.

Against the Fantasy of Autonomy
The Island can be read as pushing against autonomy as a credible outcome. The island—often imagined as a refuge or self-sustaining unit—appears instead as a structure shaped by extraction and exclusion. This reading resonates within a technological landscape that frequently markets independence through platforms, automation, and smart infrastructures, even as risk and responsibility are redistributed downward.

At Osservatorio, the elevated viewpoint functions as a spatial condition through which this logic becomes legible. From above, the Galleria’s flows appear orderly and contained, while the distance of that perspective underscores how decisions made at one level shape experiences at another. The spatial arrangement invites consideration of who sees, who decides, and who absorbs the consequences of systemic design choices. The work’s restraint is evident in what it withholds. The Island does not propose alternatives or solutions, nor does it frame oversight as empowerment. Instead, it concentrates attention on the conditions through which separation is produced and normalized.

Creative Practice Under Conditions of Infrastructure
The Island can be understood as operating less as a statement than as a working example of how contemporary practice engages systems without visualizing or mastering them. The installation does not present speculative interfaces, data dashboards, or forms of performative transparency. Instead, spatial arrangement, restricted sightlines, and institutional framing shape how the work is encountered—reflecting the partial, indirect ways infrastructures are typically experienced.

This is especially legible in how agency is structured. Viewers are not invited to interact, optimize, or intervene; their role is largely limited to navigation and interpretation. Control remains present but displaced, embedded in architecture, access protocols, and perspective rather than explicit commands—a condition that echoes how many users encounter platforms, logistical networks, and governance systems. The Island offers a reference point rather than a solution. It shows how design can register systemic conditions without translating them into tools or outcomes. The exhibition functions as a kind of infrastructural cross-section, briefly exposing how visibility, elevation, and enclosure organize experience.

Situated within Osservatorio—a cultural venue embedded in a commercial landmark—the project brings cultural and economic infrastructures into the same frame. Rather than resolving this overlap, the work holds it in place, allowing institutional conditions to remain visible. In this sense, The Island aligns with a growing body of practice in which design is treated less as interface-making than as the arrangement of conditions under which interfaces, decisions, and futures take shape.

As artificial intelligence saturates design discourse, computation is still too often framed as something that happens inside software: code executing on silicon, data processed by remote servers, intelligence abstracted from matter. Yet a parallel lineage in design research challenges that assumption. Across architecture, materials science, and fabrication, designers are working with materials that sense, respond, adapt, and reorganize—performing forms of computation without electronics, sensors, or centralized control. This shift reframes intelligence not as an overlay added through software, but as something embedded in physical behavior itself. Under the right conditions, materials can compute.

Material behavior as distributed computation
At the heart of this approach is a consequential proposition: computation does not have to be confined to software or electronic control systems. In material-centered design research, behavior can be produced through distributed physical processes, where transformation is driven by how materials are structured rather than by step-by-step digital instruction. Instead of executing code, these systems rely on geometry, composition, and internal forces to produce predictable responses. Behavior emerges through interaction, not command.

The work of the MIT Self-Assembly Lab, founded and led by Skylar Tibbits, exemplifies this logic. The lab investigates self-assembly and programmable materials—systems in which components can autonomously fold, aggregate, or change shape when exposed to external stimuli. Across projects involving self-folding sheets and shape-changing composites, “programming” is achieved through material gradients, residual stresses, and environmental triggers such as heat or moisture, rather than through sensors or active control. Self-assembly is framed as a process driven by local interactions, where order emerges without centralized coordination.

What is at stake in this work is less technological novelty than a reallocation of control. Rather than relying on external computation to determine behavior in real time, decision-making is embedded into the material system itself. The material does not execute instructions in the conventional computational sense; it responds to conditions it has been designed to register. Folding, bending, swelling, or aggregation function as physically encoded responses, unfolding in parallel across a system and reducing reliance on continuous energy input or mechanical mediation. This approach aligns with research in morphogenesis and systems biology, where complex form arises from local interactions rather than top-down specification. Designers working with material computation similarly define constraints, thresholds, and material tendencies, allowing outcomes to emerge through interaction with environmental forces. Intelligence, in this framing, is not concentrated in an algorithm or controller, but distributed across a system’s material organization.

Environmental responsiveness vs. digital control
Material computation becomes especially legible when contrasted with digitally controlled responsiveness. Many smart buildings, adaptive façades, and responsive products rely on networks of sensors, actuators, and control software—technical layers that require continuous energy supply, calibration, and maintenance. Environmental data is captured, processed, and translated into action through centralized or semi-centralized control logic.

Materially responsive systems operate differently. Certain materials respond directly to environmental conditions through their physical properties. Hygroscopic wood expands or contracts with humidity. Bimetallic strips bend predictably with temperature variation. Fiber-based composites can stiffen, relax, or reorient under load depending on internal structure. These responses occur as continuous physical processes rather than discrete computational decisions, without relying on data capture, digital thresholds, or external controllers.

Research led by Achim Menges at the Institute for Computational Design (ICD), University of Stuttgart, has been particularly influential in advancing this approach within architecture. Through work on wood and fiber composite systems, Menges and his collaborators have shown how material behavior itself can act as a design driver, enabling structures to adapt to environmental conditions without motors, sensors, or mechanical actuation. Performance emerges from how materials are organized and fabricated, rather than from real-time digital control.

The implications extend beyond technical performance. Digitally controlled systems embed governance into software and infrastructure through decisions about which data matters, how thresholds are defined, and which responses are authorized. Materially responsive systems, by contrast, operate through embedded physical logic that is legible at the scale of the material itself and difficult to modify remotely. While they trade fine-grained precision for coarser forms of adaptation, they often favor robustness, continuity, and long-term stability—qualities increasingly relevant for designers concerned with autonomy, transparency, and durability.

Sustainability implications of non-electronic intelligence
The environmental stakes of material computation are difficult to ignore. As AI-driven systems expand, so does their dependence on energy-intensive infrastructure, including data centers, sensor networks, rare earth materials, and short-lived electronic components. In this model, intelligence is tightly coupled to extractive supply chains and ongoing energy consumption.

Materially based forms of intelligence offer an alternative. Systems that rely on physical behavior rather than electronics can operate without embedded sensors, firmware updates, or network connectivity, often drawing energy directly from environmental conditions such as humidity, temperature, gravity, or load. Because behavior is governed by material properties rather than software, degradation typically occurs through wear or environmental exposure rather than sudden technical failure, shifting maintenance toward material stewardship. At building and urban scales, this enables different design strategies: adaptive shading elements that respond to climate without motors, structural components that redistribute stress through geometry, and infrastructure designed to accommodate seasonal variation passively rather than through constant optimization.

Institutions such as the Institute for Advanced Architecture of Catalonia (IAAC) have been instrumental in translating these ideas into applied research and education. IAAC’s work on bio-based materials, climate-responsive systems, and low-energy fabrication positions material intelligence as a practical design strategy rather than a speculative abstraction. Here, computation is not removed from material reality but embedded within it, aligning performance with ecological constraint. This shift does not imply a rejection of digital tools. Simulation, modeling, and digital fabrication remain essential for designing material systems whose behavior is complex and nonlinear. The distinction lies in where intelligence ultimately operates: digital tools define conditions and parameters, while the resolution of behavior is delegated to material systems rather than continuously managed through software.

Designing with behavior, not control
As materially responsive systems gain traction, they prompt a reassessment of long-standing assumptions about authorship and control. When behavior emerges through interaction rather than direct instruction, the designer’s role shifts from specifying final form to shaping conditions and tendencies. Performance is evaluated less by precision than by adaptability.

As ecological pressures intensify and technological systems grow more opaque, interest in intelligence that is visible, local, and materially grounded continues to grow. Materially responsive systems do not promise total control. Instead, they demonstrate how design can operate within uncertainty. The most consequential forms of design intelligence may not run entirely on software, but through material processes that swell, bend, crack, and recover—quietly computing in plain sight.

At ArtScience Museum, The Drone Shepherd extends Liam Young’s long-running investigation into planetary systems, automation, and environmental governance. Presented as a VR graphic novel, the work reframes Young’s speculative project Planet City through a tighter narrative lens—one focused less on the scale of the city than on the mechanisms that manage it. Rather than presenting technological salvation or dystopian collapse, the work constructs an encounter with how control, care, and infrastructure converge at planetary scale.

Young’s practice has consistently treated architecture as a medium for storytelling about systems rather than buildings. The Drone Shepherd continues that trajectory, using immersive media to render abstract processes—ecological monitoring, autonomous decision-making, remote stewardship—into spatially legible scenes. The work is speculative, but its materials are familiar: drones, sensors, automated oversight, and data-driven management. What changes is how these elements are arranged and experienced.

From Planet City to the Operational Layer
Planet City proposed a radical spatial consolidation: all humans living in a single hyper-dense metropolis to allow the rest of the planet to recover. The Drone Shepherd does not revisit that premise directly. Instead, it occupies the operational layer beneath it—the everyday systems required to make such a world function. The VR narrative unfolds in fragments, presenting landscapes shaped by drone surveillance and automated environmental control.

This shift in focus is subtle but consequential. Rather than debating the feasibility of a planetary city, the work examines the logic that governs its aftermath. Drones patrol rewilded zones, manage ecological thresholds, and enforce boundaries. The environment is no longer untouched nature nor traditional urban space, but a managed terrain optimized through continuous monitoring. The viewer’s position inside the work reinforces this emphasis. You are not an external observer assessing a master plan. You move through scenes that suggest procedures rather than places, encountering infrastructure as lived condition. The speculative question becomes procedural rather than monumental: what does planetary care look like when it is executed by autonomous systems?

Graphic Novel Logic in Virtual Space
Formally, The Drone Shepherd resists the dominant aesthetic of VR as photoreal simulation. Young adopts the visual grammar of graphic novels—bold line work, flattened depth, and segmented compositions—translated into three-dimensional space. Scenes read as panels you inhabit rather than worlds you explore freely. By avoiding realism, the work foregrounds interpretation over immersion. The work encourages the reading of systems rather than the simulation of reality. Movement is deliberate, narrative pacing controlled, and visual density calibrated to guide attention rather than overwhelm it.

VR here functions as an interface for systems thinking. The medium is used to spatialize relationships—between drones and territory, governance and ecology—without collapsing into spectacle. The experience feels closer to navigating a spatial essay than inhabiting a virtual environment, aligning with Young’s broader interest in architecture as a narrative and critical tool.

Automation as Environmental Stewardship
The figure of the “shepherd” is deliberately ambiguous. In The Drone Shepherd, care is distributed across autonomous agents that monitor, decide, and intervene without human presence. Environmental stewardship is rendered as a continuous, automated process—less a moral choice than an operational mandate. This framing echoes contemporary developments in climate technology: AI-driven monitoring systems, automated conservation efforts, and remote sensing platforms designed to manage ecological risk. Young’s work does not dramatize these technologies as villains or saviors. Instead, it presents their logic as infrastructural fact.

Scenes emphasize thresholds, boundaries, and maintenance rather than growth or progress. The environment appears stabilized but tightly regulated. Life persists, but within parameters defined elsewhere. The viewer encounters a world where governance is embedded in code and logistics, operating persistently and quietly. The political tension emerges through accumulation rather than declaration. As scenes unfold, it becomes clear that planetary repair, when delegated to machines, also redefines agency. Decisions are made continuously, while deliberative processes remain largely invisible. The work invites viewers to notice this absence rather than resolving it.

Speculation as Design Method
The Drone Shepherd functions as a case study in how speculative practice can interrogate real systems without proposing solutions. The project does not offer a model to adopt or reject. It constructs a scenario precise enough to feel operational, yet open enough to remain contested.

Presented within ArtScience Museum’s programmatic focus on intersections between art, science, and technology, the work is notable for its restraint. There are no dashboards, no interactive controls, no invitation to optimize the system. The viewer’s role is interpretive, not managerial. This choice aligns with a growing recognition that design’s influence increasingly operates at the level of infrastructure. Decisions about automation, thresholds, and governance are rarely visible in finished artifacts, yet they shape behavior and possibility at scale. The Drone Shepherd makes these decisions spatially perceptible without foregrounding aesthetic spectacle. Young’s use of VR and graphic narrative demonstrates how emerging media can support critical design inquiry rather than dilute it. By translating planetary governance into a sequence of inhabitable scenes, the work reframes speculation as a method for examining power embedded in systems.

The Drone Shepherd does not predict a future. It constructs a lens through which existing trajectories—automation, environmental management, infrastructural governance—can be examined with clarity. Its value lies in this precision: a reminder that the most consequential design work often unfolds quietly, embedded in the systems that decide how worlds are maintained.

In a newly released report, the U.S. Copyright Office has clarified how U.S. copyright law applies to works created using generative artificial intelligence, reinforcing that copyright protection requires sufficient human creative control over the expressive elements of the work.

Published in January 2025 as Part 2 of the Office’s ongoing study on artificial intelligence, the report focuses specifically on the copyrightability of works that incorporate AI-generated material. Its central conclusion is consistent with prior Office decisions: content generated entirely by an AI system, without sufficient human creative input, is not eligible for copyright protection under existing law.

Crucially for designers, artists, and creative technologists, the report draws a distinction between using AI as a tool and delegating authorship to an automated system. The Office explains that simply entering prompts or selecting from AI-generated outputs does not, by itself, meet the threshold for human authorship. To qualify for protection, a human must exercise creative control over the expressive elements of the work—such as through selection, arrangement, modification, or integration of AI-generated material into a broader human-authored composition.

The guidance does not introduce new law, but it consolidates and formalizes principles that have emerged through recent registration decisions and public consultations. For creative professionals working with generative systems, the implications are practical rather than theoretical: copyright claims will depend on how AI tools are embedded within a broader creative process, and how clearly human contribution can be demonstrated.

The Office also emphasizes that its conclusions are grounded in current U.S. copyright statutes, leaving broader policy questions—such as whether the law should evolve to address AI authorship—to Congress. Additional parts of the Office’s AI study, including analysis of training data and fair use, are expected to follow. For a creative industry increasingly shaped by generative systems, the report underscores a clear message: AI may accelerate production, but authorship—and the legal rights attached to it—still hinges on human agency.

As climate volatility, resource constraints, and infrastructural fragility intensify, many artists and architects have already moved beyond treating computation as a neutral design aid. Across architecture, landscape, and spatial research, algorithms are routinely used not just for optimization or representation, but to model territorial processes, negotiate competing land uses, and structure infrastructures that respond to environmental change over time.

Often discussed under the rubric of algorithmic ecologies, this body of work positions computation as a medium of spatial governance as much as form-making. Research studios and experimental practices increasingly integrate planetary-scale datasets, environmental modeling, and systems thinking into design workflows. The result is an approach that treats cities, landscapes, and infrastructures as adaptive systems—assemblages shaped by feedback loops, thresholds, and uncertainty rather than fixed master plans or static end states.

From Computational Design to Territorial Intelligence
Early computational architecture largely emphasized geometry, performance optimization, and formal complexity. Contemporary algorithmic practices, by contrast, increasingly operate beyond the scale of the building envelope. They engage territorial conditions by integrating climate models, satellite imagery, logistical systems, and political boundaries into spatial analysis and design research. This approach is exemplified by Territorial Agency, founded by John Palmesino and Ann-Sofi Rönnskog. The practice is known for large-scale spatial investigations that combine cartography, environmental data, and political analysis. Rather than proposing discrete architectural objects, Territorial Agency develops visual and computational frameworks that examine how energy systems, climate dynamics, and geopolitical forces interact across regions.

Their work illustrates a defining characteristic of algorithmic ecologies: design as a mode of inquiry. Computational tools are used to surface relationships—between extraction and habitation, infrastructure and ecology—rather than to reduce complex territorial systems to singular solutions. For architects and designers, this reflects a shift away from deterministic planning toward scenario-based and probabilistic spatial thinking.

Speculation as Infrastructure: Design Earth
If Territorial Agency emphasizes analytical clarity, Design Earth foregrounds speculation as a critical design method. Led by Rania Ghosn and El Hadi Jazairy, the practice produces projects that combine environmental data, geopolitical research, and narrative visualization to examine how futures are framed and governed.

Design Earth’s work frequently employs computational mapping and modeling not to predict outcomes, but to construct alternative imaginaries. Their projects visualize offshore energy fields, planetary extraction zones, and post-carbon infrastructures as contested ecological systems. In this context, algorithms function less as tools of efficiency and more as instruments for critical storytelling, making visible the assumptions embedded in infrastructural decision-making. This speculative dimension is central to algorithmic ecologies. By rendering often invisible systems legible—whether atmospheric flows or logistical corridors—designers create space for political and ethical debate. Computation becomes a means of intervening upstream in how infrastructural futures are conceived.

Adaptive Infrastructures and Urban Feedback Loops
While speculative practices shape discourse, parallel research initiatives are translating algorithmic thinking into operational research frameworks. Future Cities Laboratory, a research initiative of the Singapore-ETH Centre, investigates how cities can adapt to environmental stress through integrated modeling of urban systems. The lab’s work spans climate-responsive urban design, mobility systems, and data-driven planning tools. Crucially, it treats cities as evolving systems shaped by continuous feedback—between climate, policy, behavior, and infrastructure. Computational models are used to simulate scenarios, test interventions, and evaluate trade-offs rather than to enforce static outcomes.

This orientation reflects a broader trend: adaptive infrastructure design. Instead of building for a single predicted future, algorithmic ecologies emphasize flexibility, redundancy, and responsiveness. Sensors, simulations, and data platforms are embedded into urban systems to enable adjustment over time—whether in flood management, energy distribution, or mobility networks.

Why Algorithmic Ecologies Matter
The significance of algorithmic ecologies lies less in technological novelty than in how creative practice engages power. Algorithms increasingly mediate access to land, resources, and mobility, and when designers participate in shaping these systems, they intervene—often indirectly—in forms of governance.

Artists and architects working in this space are also reconfiguring authorship. Projects can be less of a finished object than a set of conditions: rules, datasets, models, and scenarios that structure future decisions. This reflects a broader recognition that design operates as infrastructure, shaping behavior, priorities, and possibilities through everyday systems rather than isolated artifacts.

Algorithmic ecologies provide a framework for engaging complexity without retreating into abstraction. They demonstrate how computation can support ecological thinking at scale while remaining attentive to political and cultural contexts. As climate pressures intensify and urban systems face increasing uncertainty, these practices point toward a recalibration of design’s role—not as a standalone problem-solver, but as a critical participant in shaping adaptive and contested futures.

Working within the constraints of a highly regulated industrial environment defined by hygiene protocols, security systems, and advanced manufacturing standards, Blandet sourced discarded syringe components directly from the facility’s waste streams. These high-precision plastic elements—normally engineered for laboratory accuracy—were reassembled into two functional lighting objects.

Interaction is central to the work. The syringes act as dimming mechanisms: pressing them modulates light intensity. A calibrated medical gesture becomes a tactile lighting control. Through this translation, light is treated as a measurable and compressible medium, conceptually aligned with fluid handling technologies.

The project also foregrounds questions of material value and context. Components engineered for global pharmaceutical supply chains are redirected into design objects with distinct economic and cultural positioning. Their availability is strictly site-specific; the materials are accessible only within the industrial ecosystem that produces them.

Fluid Dynamics was presented by Faber (Timișoara) as part of the exhibition Design Signals, curated by Martina Muzi at Alcova, Milan, during Salone del Mobile 2025.

Project Info
Artist: Théophile Blandet
Year: 2025
Exhibition: Design Signals, Alcova, Milan (Salone del Mobile 2025)
Presented by: Faber, Timișoara
Image Credit: Eugenio Consentino; Alex Todirică
Video: Alex Todirică

Across architecture, materials research, and experimental manufacturing, living materials have transitioned from lab-scale experiments to early-stage deployment. Mycelium composites, bacterial cellulose, and bio-reactive facades now function as test cases for a broader shift in how materials are conceived and managed. Their relevance for designers lies less in biological novelty than in the way these systems redistribute agency: when materials grow or adapt, questions of sustainability, authorship, and responsibility become operational rather than abstract.

In this context, sustainability can no longer be treated as a checklist of lower emissions or recycled inputs. Biodesign instead frames materials as active participants within ecological systems, prompting more fundamental questions. Who is responsible for a material’s lifecycle when it is partially alive? How should designers intervene—or deliberately refrain—from controlling biological processes? And what does authorship mean when outcomes are co-produced with living systems?

Biofabrication as a Design Method, Not a Material Substitute
Biofabrication is often mischaracterized as a greener substitute for conventional manufacturing. In practice, it is better understood as a methodological shift. Techniques such as growing mycelium composites, cultivating bacterial cellulose, or producing microbial pigments replace extraction and machining with cultivation, environmental control, and care.

Companies like Ecovative have demonstrated that mycelium can be grown into structural packaging, insulation, and acoustic panels using agricultural waste as feedstock. While these materials are compostable and relatively low-energy to produce, their more consequential impact is conceptual: designers must account for growth rates, humidity, contamination risk, and material decay. Control becomes probabilistic rather than absolute.

Similarly, studios such as Modern Synthesis treat bacterial cellulose not as a fixed output but as a responsive system shaped by tension, scaffolding, and environmental conditions. Design decisions occur upstream—through constraints, protocols, and feedback loops—rather than through post-processing alone. This challenges industrial assumptions about repeatability and standardization, foregrounding variability and responsiveness as design parameters.

Living Materials and the Redistribution of Authorship
When materials grow, authorship disperses. Designers define parameters, but outcomes emerge through biological processes that introduce variability and constraint. This redistribution of control is not a flaw but a defining characteristic of living materials, requiring designers to relinquish assumptions of total mastery in favor of collaboration with non-human systems. A chair grown from mycelium or a textile cultivated from microbes is not inert matter shaped solely by human intention. It is the product of metabolic labor performed by organisms. While these systems are not sentient, they operate according to their own logics, which can be disrupted by over-optimization or extractive timelines.

This reframing aligns with broader shifts in systems thinking and post-human design theory. Authorship becomes infrastructural: the designer’s role is to set conditions, define thresholds, and anticipate downstream effects. Responsibility, in turn, extends beyond aesthetics to include care, maintenance, and end-of-life pathways.

From Biofabrication to Bio-Integrated Energy Systems
Living materials are not limited to objects and surfaces. They are increasingly embedded in architectural and energy infrastructures. One of the most cited examples is the algae-powered facade developed for the BIQ House in Hamburg, which integrates bioreactor panels into a building envelope to generate biomass and thermal energy. Developed by Arup in collaboration with SSC Strategic Science Consult, the SolarLeaf system demonstrates how biological processes can be integrated into urban energy systems. Algae growth within the facade responds to sunlight, producing heat and biomass while also acting as dynamic shading.

Crucially, this is not a decorative application of “green” aesthetics. It is a systems-level intervention that links material behavior, energy production, and environmental performance. Yet it also exposes ethical tensions: bio-integrated systems require ongoing monitoring, nutrient inputs, and technical oversight. Sustainability here is not passive; it is operational and labor-intensive.

Ethics Beyond Sustainability Metrics
The ethics of biodesign cannot be reduced to carbon accounting or biodegradability claims. While many living materials offer environmental advantages, their true ethical challenge lies in how they reconfigure responsibility. Designers must consider not only how materials are made, but how they are maintained, governed, and disposed of.

There is also a risk of bio-solutionism—using living materials as symbolic fixes for systemic environmental problems. A mycelium wall panel does little if embedded in a supply chain built on overproduction and short-term use. Ethical biodesign requires alignment between material innovation and broader economic and cultural shifts toward maintenance, repair, and sufficiency. Living materials demand longer timelines, interdisciplinary collaboration, and transparency about uncertainty. They also invite designers to think politically: about who bears the cost of care, who benefits from regenerative systems, and how biological labor is valued within capitalist frameworks.

Living and regenerative materials are reshaping design not by offering a single solution to sustainability, but by forcing a reconsideration of how design intervenes in the world. As biofabrication and bio-integrated systems move from research to deployment, the central question is no longer whether these materials work. It is whether designers are willing to accept the ethical responsibilities that come with designing alongside life.

Apple’s latest expansion of Creator Studio has prompted predictable comparisons to Adobe. But framing the tool as a potential “Adobe killer” misunderstands what is actually shifting inside creative industries. Apple is not competing at the level of production. It is consolidating control at the level of circulation.

Adobe’s Creative Suite remains the dominant environment where creative labor takes form. Photoshop, Illustrator, Premiere Pro, and After Effects define how images, video, and motion are made, revised, and finalized. These tools govern craft, technique, and authorship through file formats, interfaces, and increasingly, AI-assisted workflows. Apple’s Creator Studio operates elsewhere. Accessible primarily through Apple Music for Artists, it provides analytics, campaign attribution, and visibility into how work performs once it enters Apple’s ecosystem. Streams, listeners, geographic reach, playlist placement, and marketing pathways are rendered legible—but only in aggregate, and only within Apple’s terms.

This is not redundancy; it is layering.

Creators still rely on Adobe to produce work, but they increasingly adapt that work to the metrics, recommendation systems, and editorial pipelines that Apple controls. Neither platform needs to replace the other to exert power. Together, they structure the full lifecycle of creative labor—from production to visibility.

Creative power is no longer centralized in a single tool or platform. It is distributed across overlapping control layers: software that governs how work is made, and platforms that govern how work circulates, performs, and endures. For designers, artists, and creative technologists, authorship is shaped as much by analytics dashboards and distribution constraints as by brushes, timelines, or prompts. Understanding this layered governance is now a core professional skill.Apple’s Creator Studio does not threaten Adobe’s dominance in making. It reinforces a different truth: creative labor today is negotiated across multiple systems, each quietly defining what counts as success—without ever needing to eliminate the others.

As AI systems increasingly mediate access to information, services, and decision-making, trust has become a design problem—not a branding exercise. For interaction designers working with machine learning systems, transparency is no longer limited to explainability dashboards or model cards. It is expressed through interface choices, interaction flows, defaults, and the boundaries placed on what systems are allowed to do.

Across research, practice, and platform design, a growing body of work treats trust not as user compliance but as a negotiated relationship shaped by power, legibility, and accountability. Three strands of work—critical data design, participatory HCI, and applied industry research—offer concrete models for how ethical commitments can be embedded directly into AI-driven systems.

Designing Power, Not Just Interfaces
Much of the contemporary discussion around trustworthy AI focuses on technical transparency: opening black boxes, documenting datasets, or publishing system limitations. While necessary, these moves often overlook how power is exercised at the interface layer—where users encounter, interpret, and respond to AI systems in practice. Designer and researcher Catherine D’Ignazio has consistently argued that transparency without structural accountability can reinforce existing inequalities. Through projects such as Data Feminism (co-authored with Lauren F. Klein) and her work at MIT’s Data + Feminism Lab, D’Ignazio frames design as a site where values are operationalized, not merely communicated.

From this perspective, ethical AI design is less about revealing how a model works and more about deciding who benefits, who is harmed, and who has agency within a system. Interface decisions—what data is shown, what is hidden, what actions are possible, and which are blocked—become mechanisms of governance. Transparency, then, is inseparable from politics. This shifts the role of UX from usability optimization to power-aware mediation. Design choices actively shape how AI authority is perceived, challenged, or deferred to.

Participation as an Ethical Boundary
If power is encoded through design, participation becomes a primary method for rebalancing it. HCI researcher Christina Harrington has developed participatory design frameworks that center communities historically marginalized by technological systems, particularly Black and disabled communities.

Harrington’s work emphasizes that ethical boundaries cannot be retrofitted after deployment. They must be established upstream through co-design processes that surface lived experience, contextual knowledge, and alternative value systems. In AI-driven interfaces, this often means designing constraints rather than capabilities—limiting automation where it undermines human judgment, or foregrounding uncertainty instead of confidence.

Importantly, Harrington’s research demonstrates that transparency is not universal. What is considered “clear” or “intuitive” varies across cultural and social contexts. Trustworthy systems therefore require localized forms of legibility, shaped through sustained engagement rather than generalized personas. This reframes transparency as relational. Ethical UX is not achieved by exposing more information, but by aligning system behavior with the expectations and needs of the people most affected by it.

Translating Ethics into Product Systems
While critical and academic frameworks define the stakes, their influence on real-world platforms depends on how effectively they are translated into product development processes. This translation is the focus of Google’s People + AI Research initiative, Google PAIR, which has developed widely used tools such as the PAIR Guidebook, model transparency practices, and human-centered evaluation frameworks. Google PAIR’s contribution is not a singular ethical stance, but a set of operational methods for embedding human values into AI product teams. These include design checklists for responsible AI, participatory research protocols, and guidance on communicating system uncertainty through interface design.

Crucially, PAIR’s work treats UX designers as ethical agents within AI pipelines—not as downstream stylists. Designers are positioned to define system boundaries: deciding when automation should defer to human oversight, how confidence is communicated, and how users are invited to contest or correct system outputs. While corporate constraints inevitably shape what is possible, PAIR’s influence illustrates how ethics can move from abstract principles into repeatable design practices. Trust, in this model, is maintained not by claims of objectivity but by consistent, legible system behavior over time.

From Transparency to Accountability
Taken together, these approaches suggest that transparency alone is insufficient. Trustworthy AI systems require accountability structures that are experienced through interaction—not hidden in documentation.

The implication is clear: the politics of AI are increasingly negotiated at the UX layer. Interaction designers are not merely translating system logic for users; they are actively shaping how authority, agency, and responsibility are distributed within AI-mediated environments. Ethical boundaries—what a system can infer, recommend, automate, or refuse—are design decisions. Transparency becomes meaningful only when paired with participation, constraint, and the possibility of refusal. As AI systems continue to scale, trust will not be granted by default. It will be designed, tested, contested, and redesigned—one interface decision at a time.

While desktop 3D printer hardware has seen few meaningful launches in recent months, innovation in additive manufacturing is continuing elsewhere: in materials. 3D bamboo filament, including products such as Fibrolon® Bamboo developed by FILOALFA, illustrates how material development is increasingly driving differentiation in a maturing fabrication ecosystem.

Bamboo-filled filaments typically combine PLA with finely processed bamboo fibers, producing prints with a matte, fibrous surface and subtle tonal variation. These materials are not new, but their growing visibility reflects a broader shift away from machine-centric innovation toward material intelligence—how composition, texture, and behavior influence design outcomes. As printer performance stabilizes around speed, enclosure, and automation, designers are turning their attention to what comes out of the nozzle rather than how fast it moves. Bamboo filaments deliberately resist the smooth, uniform finish associated with industrial plastics. Layer lines remain visible, fibers interrupt surface continuity, and variation becomes part of the object rather than a defect to be eliminated.

For architects, industrial designers, and researchers, this matters. Bamboo filaments are being used for architectural models, exhibition components, and speculative prototypes where material legibility is as important as geometry. The filament’s visual and tactile qualities communicate process and constraint, aligning with post-digital approaches that foreground making over polish.

Critically, bamboo-filled PLA should not be overstated as a sustainability solution. While bamboo is a renewable resource, the base polymer remains a bioplastic, and lifecycle impacts depend on sourcing, processing, and disposal. The significance of these materials is cultural and methodological rather than purely environmental. In a period where 3D printing hardware innovation has slowed, bamboo filament functions as a material signal - additive manufacturing is entering a phase where expression, texture, and material choice are primary sites of experimentation. For many, this may be more consequential than another incremental machine upgrade.

Speculative design has matured beyond the gallery-bound provocation. Increasingly, it is being operationalized as a way to prototype governance itself—treating design not as an outcome, but as a protocol: a set of rules, assumptions, feedback loops, and decision logics that shape how futures are negotiated before they are fixed. In this framing, designers are not predicting what comes next. They are constructing systems that can hold uncertainty, surface trade-offs, and make values legible while futures are still unwritten.

Across experimental studios and civic labs, this shift reframes design as a method for institutional inquiry. Rather than producing artifacts, these practices prototype conditions—testing how policies might behave, how publics might respond, and how infrastructures could be reconfigured when social, environmental, and technological variables collide. Three practices in particular demonstrate how speculative protocols are moving from theory into applied civic experimentation.

Dark Matter Labs: Designing Institutions as Mutable Systems
Dark Matter Labs approaches design as an instrument for institutional transformation rather than aesthetic production. The studio works at the scale of governance, finance, land, and law—domains traditionally resistant to design intervention. Its projects treat institutions as dynamic systems that can be reprogrammed through alternative logics of value, ownership, and accountability.

Rather than proposing finished policy solutions, Dark Matter Labs constructs experimental frameworks that allow new forms of decision-making to be tested in real contexts. This includes rethinking land stewardship models, exploring post-growth economic mechanisms, and prototyping governance structures that respond to ecological limits. Design, here, functions as a protocol for inquiry: mapping incentives, identifying systemic lock-ins, and revealing how power circulates through institutional architectures.

What distinguishes this work is its refusal of speculative distance. The studio embeds within live policy conversations, working alongside governments, civic organizations, and communities. Speculation becomes actionable not because it predicts outcomes, but because it clarifies choices. By making assumptions explicit and systems visible, Dark Matter Labs uses design to create space for institutional imagination grounded in material constraints.

Extrapolation Factory: Futures as Participatory Governance Tools
Founded by Elliott P. Montgomery and collaborators, Extrapolation Factory is often associated with provocative future artifacts. Yet its deeper contribution lies in how those artifacts operate as civic interfaces. The studio’s work uses speculative scenarios to engage publics, policymakers, and institutions in conversations about long-term consequences that are otherwise difficult to address.

Extrapolation Factory’s projects are structured as experiential systems: exhibitions, workshops, and public engagements that simulate alternative regulatory, environmental, or technological conditions. These are not predictions, but carefully constructed extrapolations from existing trends. By extending present-day logics into plausible futures, the studio creates shared reference points for debate.

In governance contexts, this approach functions as a protocol for collective sense-making. Participants are invited to inhabit futures shaped by today’s policy decisions, making abstract issues tangible and emotionally legible. The value lies less in the speculative object than in the process it enables—one that distributes authorship of the future across diverse stakeholders. Design becomes a mediator between expertise and public imagination, allowing governance to be rehearsed before it is enacted.

Rachel Coldicutt: Embedding Design Protocols in Civic Technology
Rachel Coldicutt operates at the intersection of digital technology, public institutions, and social outcomes. Her work emphasizes how design protocols—standards, processes, and decision frameworks—shape the ethical and practical impact of civic technologies. Rather than treating technology as neutral infrastructure, she foregrounds how design choices encode values into public systems.

Coldicutt’s approach is notably pragmatic. Working with governments, funders, and cultural institutions, she focuses on improving how technologies are commissioned, evaluated, and governed. This includes advocating for inclusive design practices, transparent evaluation metrics, and accountability mechanisms that extend beyond deployment. Speculation, in this context, is not futuristic storytelling but anticipatory governance: asking how systems might fail, exclude, or produce unintended consequences over time.

By embedding design thinking into procurement and policy processes, Coldicutt demonstrates how speculative protocols can operate quietly but effectively within institutions. The future is shaped not by grand visions, but by everyday decisions about data, accessibility, and responsibility. Design becomes a means of institutional self-reflection—an ongoing practice rather than a one-off intervention.

From Speculation to Stewardship
Taken together, these practices signal a broader shift in how speculative design is understood and applied. The focus is no longer on imagining distant futures, but on designing systems capable of adapting to uncertainty. Protocols replace predictions. Governance becomes a design space. This reframing expands the field’s responsibility. Working at the level of systems requires fluency in policy, economics, and institutional behavior, as well as a willingness to engage with slow, complex change. It also demands rigor: speculative protocols must be testable, legible, and accountable if they are to influence real-world decision-making.

As climate instability, technological acceleration, and social fragmentation continue to challenge existing institutions, the ability to prototype governance itself may become one of design’s most consequential roles. Not as authors of the future, but as architects of the conditions under which futures can be collectively negotiated.

This week, NVIDIA formally introduced Earth-2, an AI-accelerated climate and weather modeling platform designed to run high-resolution environmental simulations at unprecedented speed and scale. Announced at the American Meteorological Society meeting, the system combines machine-learning models with physics-based simulation to generate localized forecasts, including short-term “nowcasts,” using significantly less compute than traditional numerical methods.

While Earth-2 is positioned as infrastructure for climate science, disaster preparedness, and energy planning, its implications extend well beyond meteorology. For designers, architects, and creative technologists, Earth-2 reframes simulation as an active design medium rather than a downstream analytical tool. Unlike conventional climate datasets—often static, coarse, and retrospective—Earth-2 operates as a continuously updating model of environmental behavior. This enables iterative testing of scenarios at the scale of neighborhoods, coastlines, or regions, making climate conditions a live constraint within design workflows. For practices working in parametric architecture, responsive urban systems, or environmental visualization, the platform signals a shift toward climate-informed form-finding driven by near-real-time data rather than historical averages.

Earth-2 also marks a broader evolution in NVIDIA’s strategy. Building on the logic of Omniverse and digital twins, the company is moving from simulating discrete objects and spaces toward modeling interdependent planetary systems—weather, infrastructure, energy, and risk—within shared computational environments. The result is a form of simulation that increasingly overlaps with governance, planning, and policy.

The critical question is not whether these simulations are more accurate, but how they shape decision-making. As platforms like Earth-2 become embedded in planning and design processes, they quietly establish defaults: what variables matter, what futures are modeled, and which risks are rendered visible—or ignored. Earth-2 is not just a forecasting tool. It is a signal that planetary simulation is becoming a foundational layer of contemporary design practice.

Apple’s Vision Pro is increasingly being framed less as a consumer entertainment device and more as a professional spatial computing platform. While early attention focused on immersive media and personal productivity, recent software updates, developer messaging, and partnerships signal a deliberate shift toward enterprise, engineering, and design-oriented use cases.

Since launch, Apple has emphasized Vision Pro’s ability to run familiar desktop-class applications in spatial environments. Native support for macOS screen mirroring allows professionals to scale multiple high-resolution virtual displays, positioning the headset as an alternative to multi-monitor setups rather than a replacement for traditional computing. This approach aligns Vision Pro with production workflows already common in design studios, architecture firms, and engineering teams.

Enterprise-focused collaborations reinforce this direction. Companies including Adobe and Autodesk have introduced Vision Pro-compatible tools that extend creative and technical workflows into three-dimensional space. Rather than promoting novelty, these integrations emphasize continuity: familiar software, spatially expanded. Apple has also highlighted use cases in product design review, medical visualization, and industrial training—areas where immersive context adds measurable value.

Notably absent is aggressive consumer marketing around gaming or social experiences. Instead, Apple’s messaging increasingly mirrors that of a professional workstation vendor. Vision Pro’s precise eye tracking, low-latency hand input, and high pixel density are framed as instruments for accuracy and focus, not spectacle. This positioning aligns with Apple’s broader strategy of targeting high-margin professional markets before mass adoption.

The implication is clear. Vision Pro is less about lifestyle augmentation and more about spatial computing as infrastructure—a tool for thinking, prototyping, and collaboration. As software ecosystems mature and costs eventually come down, Apple appears to be laying groundwork for spatial computing to become a normalized layer of professional production rather than a niche entertainment format. From this perspective, Vision Pro is not an experiment. It is a quiet bet that future creative work will happen in space, not just on screens.

A new research training group funded by the German Research Foundation brings together the University of Stuttgartand the University of Freiburg to investigate smart materials and responsive building envelopes. The initiative focuses on envelope systems that adapt to environmental conditions through material behavior, with particular emphasis on bio-inspired materials and compliant mechanisms as alternatives to electronically controlled adaptive systems.

The research is situated at the intersection of architecture, engineering, and materials science, addressing long-standing challenges in how buildings manage environmental performance while limiting mechanical complexity.

Rethinking the Building Envelope as an Active System
Building envelopes have historically been designed as largely static assemblies, optimized for insulation, weather protection, and visual performance. Over the past two decades, performance-driven design approaches have introduced dynamic shading systems, kinetic facades, and automated ventilation strategies. These systems have improved environmental responsiveness but often depend on electromechanical components, centralized control systems, and ongoing maintenance.

The research training group focuses on alternative approaches to adaptive envelopes, investigating systems that respond to environmental stimuli such as humidity, temperature, and solar radiation through material behavior and structural configuration. Rather than relying on motors, sensors, or software-driven control, the research explores how responsiveness can be achieved through the intrinsic properties of materials and the way they are assembled.

Drawing on bio-inspired principles, the work examines how movement and adaptation can emerge from material composition, geometry, and compliant mechanisms. In natural systems, environmental responsiveness is often achieved without centralized control, offering models for architectural components that operate passively and predictably over time.

Smart Materials Beyond Embedded Electronics
The research training group emphasizes approaches to smart materials that foreground material behavior and structural logic rather than electronically controlled systems alone. Research areas include bio-inspired composites, anisotropic materials, and hybrid structures whose performance can be influenced during design and fabrication. Computational design plays a central role in this process, serving as a tool for modeling, simulating, and predicting how material systems respond under environmental loads. These digital methods are integrated with physical workflows, including robotic fabrication and material testing, allowing researchers to evaluate behavior across multiple scales and conditions.

By exploring material-driven responsiveness, the research investigates alternatives to mechanically complex adaptive systems. These approaches are examined for their potential relevance to durability, energy use, and architectural expression, particularly in contexts where simplicity and passive performance are critical.

Training Researchers at the Intersection of Design and Science
The research training group will support approximately 20 doctoral researchers over a five-year period beginning in 2026. The program brings together doctoral projects across architecture, engineering, materials science, and biology within a structured interdisciplinary framework, reflecting the complexity of adaptive building systems.

Doctoral researchers engage with a range of scales and methods relevant to responsive envelopes, including material development, component prototyping, computational modeling, and experimental testing. As with other DFG Research Training Groups, the program includes coordinated supervision, seminars, and collaborative research formats designed to support interdisciplinary exchange and methodological rigor.

Implications for Architecture, Industry, and Design Culture
While the research training group is focused on building envelopes, the material systems under investigation relate to broader developments in adaptive structures and bio-inspired materials research. Similar approaches are being explored across architecture, construction, and adjacent design fields where lightweight, passive, and materially efficient performance strategies are increasingly relevant.
Within architectural discourse, the research aligns with ongoing interest in material-driven responsiveness and alternatives to digitally controlled adaptive systems. By combining computational design with material experimentation, the program contributes to wider discussions about how environmental performance can be addressed through structural and material behavior rather than through additional technological layers.

The European Union is moving the AI Act from legislative framework to active enforcement faster than many platforms anticipated. While the law formally entered into force in 2024, recent actions by the European Commission signal a sharper focus on early compliance, particularly around high-risk and general-purpose AI systems. For design-led AI tools, the shift marks a transition from policy interpretation to operational consequence.

Under the AI Act, prohibitions on so-called “unacceptable risk” systems—including certain forms of biometric categorization and manipulative AI—are among the earliest provisions to apply. At the same time, obligations for general-purpose AI models, including transparency requirements and technical documentation, are being clarified ahead of the Act’s full phased rollout. National authorities are now standing up supervisory bodies, while the Commission is publishing guidance intended to standardize enforcement across member states rather than leaving interpretation entirely to local regulators.

For designers and creative technologists, the immediate impact is less about consumer-facing bans and more about upstream accountability. Tools that generate images, video, text, or audio at scale will increasingly be expected to disclose synthetic content, document training data practices at a high level, and assess whether outputs could fall into regulated risk categories. Platforms operating across the EU will need to demonstrate that compliance is embedded in product design, not bolted on at launch.

The enforcement acceleration also reshapes timelines for startups and studios relying on third-party AI models. Liability and compliance obligations can extend through the supply chain, meaning creative teams may need clearer contractual assurances from model providers. As enforcement mechanisms solidify, the AI Act is becoming a design constraint as real as accessibility standards or data protection—one that will shape how creative tools are built, deployed, and governed in Europe over the next year.

Materials have historically been treated as stable substrates—chosen, specified, and deployed to perform predictably under known conditions. That assumption is increasingly inadequate. Across architecture, textiles, and industrial design, materials are now developed to behave: to respond to environmental stimuli, to change state over time, and to be tuned through computational feedback rather than fixed recipes. Beyond aesthetics, this behavior is structural. It reshapes how performance is specified, how failure is assessed, and how responsibility is distributed across design, fabrication, and maintenance.

Rather than designing a finished form, practitioners are increasingly defining behavioral envelopes—ranges of action materials can perform under varying conditions. Intelligence, in this context, does not necessarily mean embedded computation. It often resides in geometry, composition, fabrication logic, or in the workflows used to discover and optimize material properties.

Programmable textiles beyond embedded electronics
Smart textiles were historically dominated by add-on approaches—sensors stitched into fabric, electronics laminated onto garments, and conductive threads treated primarily as carriers for hardware. Current research increasingly targets the material substrate itself. At Aalto University, projects such as SuperTextil investigate conductive, bio-based textiles using scalable graphene coatings, targeting functions such as electrical heating and sensing while reducing reliance on metals and discrete electronic components. The project foregrounds research priorities including durability, washability, and compatibility with existing textile manufacturing processes.

In parallel, Aalto’s AI-yarn project applies machine learning techniques to textile material development itself. Rather than relying solely on manual trial-and-error, researchers use computational optimization to explore how variables such as coating thickness, material composition, and processing parameters influence performance outcomes. Here, “intelligence” resides not in the fabric alone, but in the development pipeline that shapes it.

Taken together, these efforts point to a broader methodological shift. Materials are no longer specified solely through fixed recipes; they are increasingly developed through data-driven iteration, with designers working alongside scientists and engineers to define target behaviors rather than static characteristics.

Self-transforming systems and programmable matter
Among the research groups shaping contemporary discourse on programmable materials is the MIT Self-Assembly Lab, led by Skylar Tibbits. The lab’s work focuses on engineered systems designed to change shape, properties, or function in response to external stimuli such as heat, moisture, or mechanical force.

Importantly, these materials are not framed as autonomous or “alive.” They are deliberately constrained systems whose behaviors are specified in advance. The central design challenge lies in determining how and when transformation occurs, and in ensuring that such changes remain legible and predictable. This requires a form of design literacy that integrates material science, computation, and fabrication logic. For architecture and product design, the implications are significant. Walls, panels, and textiles can no longer be evaluated solely through static metrics. Performance unfolds over time, making testing, simulation, and verification integral to the design process.

Fabrication workflows catch up to material complexity
As materials become more behaviorally complex, traditional construction documentation often struggles to keep pace. Soomeen Hahm Design addresses this gap by integrating computational design, robotic fabrication, and augmented reality into construction workflows.

The studio’s Steampunk Pavilion, presented at the Tallinn Architecture Biennale, employed steam-bent timber elements assembled on site with guidance from augmented reality headsets, reducing reliance on conventional two-dimensional drawings. The project was not concerned with “smart” timber, but with adaptive fabrication—using digital tools to accommodate material variation during construction.

This approach is increasingly relevant. When materials vary, bend, or behave unpredictably, rigid documentation can become a liability. Designers must instead develop systems that absorb variation while maintaining structural integrity and spatial intent.

Designing for legibility, not novelty
The growing availability of programmable materials introduces a critical concern frequently noted in design and systems discourse: opacity. As materials respond, transform, or adapt, failures can become more difficult to diagnose, repair pathways less clear, and responsibility more diffuse. For designers working with behavioral materials, legibility becomes a central requirement. Clear stimulus–response relationships, modular systems, and transparent documentation are necessary if such materials are to move beyond experimental installations into durable, maintainable environments. From this perspective, the future of programmable matter is likely to hinge less on how much intelligence is embedded in materials than on how carefully behavior is governed—whether systems remain understandable, repairable, and ethically deployed over time.

Generative AI is now embedded across creative workflows, yet the systems that enable it—cloud platforms, data pipelines, and large-scale training cycles—carry environmental and political costs that remain largely out of view. As models expand and inference becomes ubiquitous, the material footprint of computation is becoming a critical design concern. Circular data, a framework centered on minimizing, reusing, or localizing datasets and compute cycles, offers a way to rethink how creative AI systems are built and sustained. Across the field, practitioners are exposing the hidden resource flows behind digital systems and proposing alternative infrastructures that prioritize ecological and ethical responsibility.

Making AI’s Materiality Visible
Joana Moll has long challenged the assumption that digital interactions are immaterial. Her investigations into the operational complexity and energy implications of online services—such as The Hidden Life of an Amazon User, which exposes the dozens of scripts and tracking processes triggered by a simple purchase—reveal how digital activity depends on continuous resource consumption. Her broader work examines the carbon footprint of tracking and advertising infrastructures, making visible the environmental impact of data-intensive services that typically operate out of sight.

While Moll does not focus specifically on generative AI, her research on the material costs of digital systems provides a critical foundation for understanding the environmental pressures that AI development amplifies. Large-scale models rely on electricity, specialized hardware, and distributed cloud infrastructure at every phase of their lifecycle, from dataset preparation to training and inference. Recognizing these resource flows is essential for any effort to build AI systems with a reduced ecological footprint. A circular data approach builds on this awareness by shifting attention from unbounded dataset growth toward more intentional, efficient, and context-driven data practices.

Mapping the Politics of Infrastructure
If Moll foregrounds the material costs of computation, Ingrid Burrington reveals the infrastructures that make those costs possible. Her documentation of fiber-optic cables, data centers, and the territorial footprints of networked systems offers a lens through which to understand the physical architecture that AI now depends on. Burrington’s research surfaces the civic and geopolitical dimensions of the cloud: the energy grids that power data centers, the land they occupy, and the regulatory structures that govern them.

Working with generative models, this perspective reframes cloud usage as an infrastructural choice rather than a neutral technical detail. Decisions about where data is stored, which providers host training runs, and how models interface with networks all have environmental and political consequences. Circular data extends naturally from this view, emphasizing locality, transparency, and an awareness of how computational workloads circulate across global infrastructures.

Designing Alternative Pipelines
Creative technologists are also demonstrating how pipeline design itself can counter the default logic of large-scale, cloud-dependent AI. Björn Karmann’s work offers concrete examples of this shift. Project Alias places a programmable, user-controlled device over the microphones of commercial smart speakers. It continuously emits noise to block passive listening and, when activated, plays a custom wake word that allows the speaker to receive commands. The device inserts a local layer of control between users and cloud-based assistants, reducing unnecessary data transmission and challenging the assumption that continuous upstream audio is required for usability.

Karmann’s Paragraphica likewise explores alternative relationships between data and output. Instead of using photographic light capture, the system gathers geolocation and environmental data—such as place, time, or nearby points of interest—and uses those inputs to generate an image via existing pre-trained models. While these models are still built on large-scale training, the project shifts emphasis away from dataset accumulation toward contextual sensing and recombination. For creative technologists, this approach demonstrates how constrained, situational data can open new modes of experimentation without defaulting to building or retraining large models.

These practices reflect the core principles of circular data: reduce transmission, use what is already available, localize where possible, and design workflows that avoid unnecessary compute cycles.

Data Circularity as an Ethical and Political Framework
Circular data is not only a technical proposition. Joana Varon and the Brazilian organization Coding Rights expand the conversation to include the rights-based and geopolitical implications of data practices. Their work examines how data collection and AI development reflect broader patterns of extraction, often disproportionately affecting marginalized communities in the Global South. Coding Rights’ analyses of data colonialism, governance, and AI accountability highlight how training datasets—whether newly created or repeatedly reused—shape the power dynamics embedded in computational systems.

Adopting circular approaches, this raises critical questions: What forms of reuse respect community consent? How can localized datasets support autonomy rather than reinforce inequities? As models become more powerful and accessible, sustainable AI must also engage with social and cultural dimensions of data stewardship. Circular data, approached through this lens, integrates ecological responsibility with ethical governance.

Building Toward Sustainable AI Systems
As AI becomes a routine part of creative production, the environmental and infrastructural realities of computation need to become part of the design vocabulary. Circular data offers a way to reframe the lifecycle of training and inference—reducing redundancy, rethinking storage and transfer, and encouraging the use of contextual or localized data. It signals a shift from scale-driven development toward sufficiency, adaptability, and transparency.

Moll’s environmental investigations, Burrington’s infrastructural mapping, Karmann’s speculative pipeline design, and Varon’s rights-based critique together illustrate how a more sustainable AI ecosystem might emerge. Each contributes a different but complementary insight: computation has material limits, infrastructure is political, pipelines can be reimagined, and data governance matters.

The challenge, of course, is to integrate these insights into practice—designing systems that acknowledge their dependencies, minimize their footprint, and align with broader commitments to ecological and social responsibility. Circular data does not solve the environmental pressures of AI outright, but it provides a framework for building systems that consume less, reveal more, and operate with greater accountability as generative computation continues to evolve.

The aesthetics of machine vision don’t begin with neural networks; they begin with the dataset. As image models, recommendation systems, and civic algorithms weave into daily life, artists and curators increasingly treat datasets not as neutral inputs but as archives, lenses, and infrastructures of power. Gene Kogan, Hannah Redler Hawes and the Open Data Institute’s Data as Culture programme, and James Bridle approach this from different directions, yet all foreground the dataset as a site of authorship, control, and critical inquiry.

The Dataset as Medium, Not Just Fuel
For decades, “data” functioned largely as back-end substrate—an engineering concern that remained separate from the aesthetic decisions of designers and artists. That boundary has eroded. As machine learning becomes embedded in creative practice, the dataset itself has moved into view as an expressive and conceptual medium.

Artist and programmer Gene Kogan has been central to this shift. His long-running project ml4a (Machine Learning for Art) is a public collection of open-source code, tutorials, workshops, and an in-progress textbook designed to help artists work directly with machine learning systems. The project is explicitly described as “a collection of tools and educational resources which apply techniques from machine learning to arts and creativity,” and was initiated and maintained by Kogan as an open resource between 2015 and 2021.

Kogan has reinforced this pedagogical approach through institutional teaching, most notably his Machine Learning for Artists course at New York University’s Interactive Telecommunications Program (ITP). In that course—outlined in his Medium posts and course materials—artists and designers are introduced not only to neural network architectures and generative models but also to dataset curation, data preprocessing, and labeling practices. The emphasis is clear: critical engagement with AI requires understanding how training data conditions what a model can perceive, classify, or generate.

Kogan’s later research extends this argument into collaborative, decentralized systems. His project Abraham, introduced in a paper for the NeurIPS 2019 Creativity Workshop and expanded in subsequent writing, proposes an “autonomous artificial artist” built and governed by a distributed network of contributors. Abraham is conceived as a long-term, open project in which participants help design, train, and refine a generative art agent. In this model, the dataset becomes a collectively authored cultural object—its contents determining the aesthetic scope, tendencies, and biases of the artificial “artist.”

This logic reframes a foundational idea for creative AI practice: dataset authorship is aesthetic authorship. Decisions about what the machine sees—what images, languages, histories, or bodies populate its training corpus—are not neutral technical steps. They are formative aesthetic acts, shaping the system’s expressive vocabulary as directly as any gesture, material choice, or compositional decision made by a human artist.

Data as Culture: Curating Data’s Social and Civic Aesthetics
If Kogan opens up the technical stack, Hannah Redler Hawes builds institutional infrastructures around data as cultural material. As Director of the Data as Culture art programme at the Open Data Institute (ODI), she works at the intersection of contemporary art, data, and the social and ethical questions raised by digital technologies.

Data as Culture has, since 2012, commissioned and exhibited artworks that use data as an art material while interrogating its social and ethical implications. The ODI frames the programme as a way to engage diverse audiences with artists and works “that use data as an art material,” situating artistic practice alongside its broader work on open and trustworthy data ecosystems. Early commissions—documented in ODI catalogues and talks for the 2014 Data as Culture exhibition—focused explicitly on surveillance, privacy, and personal data, using pneumatic machines, satellite imagery, data-collection performances, and “knitted data discrepancies” to explore how open data and monitoring intersect with everyday life.

A key anchor for the programme’s long-term trajectory is Julie Freeman’s online work We Need Us (2014–ongoing), co-commissioned by the ODI and The Space and later exhibited with NEoN Digital Arts. The piece pulls live metadata—data about data—from the citizen science platform Zooniverse, transforming the collective activity of over a million volunteers into a dynamic environment of animated forms and sound. Rather than visualising scientific results, We Need Us focuses on the rhythms, density, and volatility of participation itself, exploring what Freeman has described as “the life of data.”

More recently, Freeman’s Allusive Protocols (2023), commissioned by Data as Culture at the ODI with support from Invisible Dust, uses kinetic, data-driven sculptures to reflect on networked infrastructure and the fragility of connectivity. The work responds directly to ODI research on power and diplomacy in data ecosystems, considering how control over protocols and networks underpins contemporary infrastructure and the distribution of power.

Across more than a decade, Data as Culture has presented over 100 works—including 27 commissions across 11 exhibitions and partnerships—and has developed research-led themes such as Copy That?, which asks how “true” the “data you” really is and how many versions of that self exist online. This is curation as infrastructural critique: datasets are treated as civic artefacts, and their aestheticisation is tied to questions of accountability, access, and governance.

James Bridle: When Datasets Become Governance
Artist and writer James Bridle pushes the aesthetics of the dataset into the domain of political technology. Known for coining the “New Aesthetic” and for books such as New Dark Age: Technology and the End of the Future (Verso, 2018) and Ways of Being (2022), Bridle traces how computation, automation, and data infrastructures reorganise perception and power.

In the solo exhibition Failing to Distinguish Between a Tractor Trailer and the Bright White Sky at NOME Gallery in Berlin (2017), Bridle took the self-driving car as a central motif. The title comes from an accident report into a fatal Tesla crash in which the vehicle’s Autopilot system failed to distinguish the white side of a tractor-trailer against a brightly lit sky—an incident widely reported as exposing limits in the car’s sensors and system design.

The exhibition gathered works that probe machine learning and machine vision: how autonomous driving systems are trained, how they “see” the world, and how those ways of seeing can fail. Bridle worked with “software and geography” to create components for his own self-driving car—an autonomous vehicle “which learns to get lost”—using freely available tools and research papers. Through installations, image series, and the video essay Gradient Ascent, which documents test drives on Mount Parnassus, the exhibition linked technical processes such as optimisation and training to questions of labour, responsibility, and the political opacity of complex systems.

Autonomous Trap 001 (2017) condenses these concerns into a single scene: a car encircled by a salt line and road markings laid out as inward-facing “no entry” signs. As Bridle’s own description notes, the work uses ground glyphs “to trap autonomous vehicles using ‘no entry’ and other glyphs”; the car, programmed to obey those markings, cannot leave the circle without breaking its encoded rules. The trap here is not magical but infrastructural—an environmental configuration that reveals how rule-bound systems can be manipulated, and how much agency is ceded to code, training regimes, and the datasets on which they depend.

In New Dark Age and Ways of Being, Bridle extends this critique beyond the gallery. Both books argue that contemporary knowledge is increasingly routed through computational systems—search engines, models, and platforms—that are trained on vast, uneven datasets and embedded in economic and political structures. Rather than treating data as neutral input, Bridle shows how data-driven infrastructures decide what is visible, what is legible as “information,” and whose experience is excluded. In that sense, datasets become epistemological tools: they shape what counts as knowledge for human institutions and for machines alike, and their blind spots often mirror the social and environmental hierarchies in which they are produced.

Curating the Machine’s Eye
Taken together, these practices point to a shared cultural shift. Kogan equips artists and designers to work directly with training data and models, treating machine learning literacy as a creative skill set. Redler Hawes and the Data as Culture programme develop public-facing contexts in which data-driven artworks are exhibited, questioned, and contextualised—from live metadata works such as We Need Us to commissions that address energy systems, connectivity, and the power dynamics of networked infrastructures. Bridle, meanwhile, shows how training data, sensor inputs, and encoded rules can function like policy in practice, influencing what systems are able to perceive, how they behave, and where their failures become visible.

For a design-savvy, technologically fluent audience, this is the real frontier of AI aesthetics. Style transfer and text-to-image prompts represent only the surface layer. The deeper work lies in curating the machine’s eye: determining which archives matter, which signals count as noise, whose worlds become legible to computation, and how those decisions are opened up to scrutiny, collaboration, and repair. The aesthetics of the dataset are, ultimately, the aesthetics of the systems we build—and the futures we enable through them.

Urban design is shifting away from a human-exclusive worldview. Climate instability, biodiversity collapse, and the spread of networked sensors are exposing the limits of cities built solely for human optimization. A new model is emerging—one where organisms, machine systems, and environmental processes become co-authors of urban space. This is not an aesthetic turn toward biomorphism; it is a strategic realignment of how infrastructure works, how it is maintained, and who it serves.

Across landscape architecture, urban technology, civic art, and architectural research, a set of practitioners is building systems that treat oysters, air particulates, seeds, sensors, and autonomous digital agents as operational components of the city. Their work signals a future in which urban resilience depends on symbiosis, feedback, and distributed intelligence rather than control and centralization.

Ecological and Atmospheric Systems as Infrastructure
Landscape architect Kate Orff and her studio SCAPE exemplify how ecological processes can function as core infrastructure. Projects like Living Breakwaters demonstrate how engineered structures can be designed for biological colonization, allowing oysters to mitigate wave energy, strengthen shorelines, and improve water quality. This model challenges the convention of static hard infrastructure by substituting it with evolving ecological systems that require long-term stewardship rather than one-time construction.

On a parallel track, Studio Roosegaarde explores how atmospheric sensing can become a public interface. The Smog Free Tower, for example, uses ionization to remove particulate matter while transforming air-quality metrics into a visible, spatial experience. Here, environmental data is not a background diagnostic but an active element that shapes public awareness and behavior. Together, these approaches illustrate how biological and technological systems can operate as co-infrastructures—each responding to conditions humans alone cannot manage.

Multispecies Participation and Networked Governance
The more-than-human city is not solely a technological or ecological construct; it is also a civic one. The collective Futurefarmers builds participatory frameworks that link everyday urban life to ecological and agricultural systems. Seed libraries, mobile research units, and collaborative workshops demonstrate how public engagement becomes a critical layer of urban metabolism. Their work shows that multispecies urbanism depends not just on environmental engineering but on shared governance, ecological literacy, and cultural practices capable of sustaining long-term ecological care.

Meanwhile, architectural research studio Space Caviar addresses the other major non-human actor embedded in cities: digital infrastructure. Their investigations into sensor networks, supply-chain logistics, algorithmic governance, and platform-based housing reveal how non-human digital systems influence spatial access and resource distribution. These systems operate alongside biological actors—shaping flows, mediating decisions, and generating new forms of urban governance.

Toward Symbiotic Urbanism
What emerges from these practices is not a single vision but a shared operational shift. Cities are becoming ecological-technological hybrids where oysters, particulate sensors, microbial systems, autonomous agents, and human communities co-produce resilience. Designing for this reality requires architects, technologists, and planners to orchestrate relationships rather than impose fixed solutions. It also demands regulatory frameworks capable of supporting living, adapting, and sensing infrastructures that exceed traditional engineering categories.

The more-than-human city takes shape through these experiments: ecological infrastructures that grow stronger through biological activity, public spaces that react to atmospheric signals, civic practices that cultivate multispecies responsibility, and architectural research that makes algorithmic systems visible. As these models scale, the city becomes not a human-only environment but a shared platform for biological, technological, and cultural interaction—one that is better equipped for the uncertainties ahead.

Apple has rolled out an updated hardware revision of its Vision Pro headset, now powered by the company’s M5 chip. The upgrade introduces a next-generation 10-core CPU, more powerful GPU, and enhanced Neural Engine under a 3 nm process — improvements that collectively yield sharper visuals, smoother performance, and accelerated AI tasks.

Among the headline specifications: Micro-OLED displays rendering 10 percent more pixels than the prior model, and an increased refresh rate of up to 120 Hz to reduce motion blur during passthrough and virtual-display use. Apple also ships the headset with a new “Dual Knit Band,” designed for improved comfort and stability. 

Yet, despite the technical leap, the reception remains cautiously measured. According to Ars Technica’s recent review, while the M5-equipped Vision Pro delivers smoother performance — especially for graphics-intensive and AI-powered spatial apps — its value proposition remains constrained by persistent issues: limited content catalog, high price point, and ergonomic trade-offs relative to lighter AR glasses. 

For designers, developers, and creative professionals, the M5 refresh improves the headset’s viability for pro workflows: real-time 3D visualization, spatial UI, and AI-assisted design tools may now run more reliably. But as the market pivots toward accessibility and portability, Vision Pro still straddles an uneasy middle ground: powerful, yet niche.

Micro-robots are moving a step closer to real-world use as several new peer-reviewed studies published in 2024 and 2025 report advances in propulsion, power delivery, and in-vivo navigation—longstanding bottlenecks that have limited the field to controlled laboratory tests.

One of the most cited developments comes from Cornell University, where optical microrobots powered by on-chip photovoltaics demonstrated improved untethered motion when steered by focused laser light. The system, developed by the Paul McEuen and Itai Cohen groups and reported in Science Robotics and Nature, shows how cell-scale robots can operate without wired power or bulky magnetic equipment. At the University of Pennsylvania, researchers in the GRASP Lab have advanced magnetically actuated micro-swimmers capable of navigating viscous fluid channels with greater stability than earlier designs. Their 2024 Science Robotics paper details improved steering precision using rotating magnetic fields—an increment that brings microrobots closer to functioning in complex biological environments.

Biomedical applications remain a driving force. Multiple groups, including teams at ETH Zurich and the Max Planck Institute for Intelligent Systems, have recently demonstrated micro-scale devices performing targeted transport tasks in animal models. None of these platforms are approved for human clinical use, but early in-vivo studies show controlled motion with minimal tissue disruption.

Environmental testing is also expanding, primarily through lab-based micro-swarm prototypes for pollutant sensing and microplastic capture. Despite momentum, major hurdles persist: power constraints, long-term biocompatibility, and the difficulty of manufacturing large, coordinated swarms. Still, the growing volume of reproducible results suggests micro-robotics is shifting from speculative prototypes toward application-driven systems with implications for medicine, environmental monitoring, and materials inspection.

Design has long borrowed from biology, but the current shift goes further than biomimicry or nature-inspired form. A growing movement toward regenerative intelligence positions living systems as collaborators in design—systems that respond, adapt, and evolve rather than remain fixed. The focus is no longer on extracting efficiency from ecological forms; it is on working with ecological processes to build resilience, feedback-driven infrastructures, and multispecies forms of cohabitation.

Across architecture, materials research, and landscape design, a set of practitioners is reframing intelligence itself as ecological: distributed, cyclical, and interdependent. Their work suggests a structural transition in design culture, one that moves away from optimization as a dominant paradigm and toward strategies that regenerate the conditions that make environments livable.

From Optimization to Ecological Feedback
Most design systems still operate around the logic of optimization—streamlining resource flows, reducing waste, or maximizing performance. These goals matter, but they often freeze environments into static configurations, leaving little capacity for adaptation. Regenerative intelligence shifts the emphasis to systems that evolve through feedback, where materials and environments adjust to external pressures rather than resist them.

This perspective aligns design with ecological realities. Forests redistribute nutrients during stress; wetlands modulate water flows; coral reefs build architectures through interaction, not control. Regenerative intelligence treats these processes as operational models rather than metaphors.

In practice, this means designing infrastructures that learn from disturbances, not simply withstand them. It also means integrating biological actors—microbes, plants, algae, or soil systems—into built environments in ways that allow for ongoing negotiation. Instead of optimizing for a narrow set of parameters, regenerative design embraces variability, succession, and transformation.
Studio Ossidiana’s practice embodies this shift. Their work often centers on material ecologies shaped by geological, biological, and cultural dynamics—projects where matter is not inert but part of a responsive system. Their installations and landscapes explore how materials behave over time, interacting with water, light, and other species rather than holding a fixed form. By foregrounding change instead of stability, their work argues for a design approach that accepts—and leverages—environmental flux.

Designing with Living Systems, Not Around Them
Regenerative intelligence also reframes the role of living systems in design. Rather than treating biological matter as a resource or aesthetic reference, it positions organisms as active participants whose agency shapes the outcome.

Anna Heringer’s architectural practice is a clear example. Her work with earthen construction demonstrates how local ecosystems, communities, and materials can collaborate to form resilient structures. Earthen buildings respond to humidity, temperature, and repair differently than industrial materials. They invite maintenance as a shared cultural practice and emphasize the social and ecological dimensions of building. Her projects show how regenerative principles—circularity, low-energy processes, and local ecological alignment—can produce architecture that strengthens rather than depletes environmental systems.

This orientation toward ecologically responsive processes is also reflected in the emerging field of biofabrication, where materials such as mycelium, algae, or bacterial cellulose become structural elements. These materials grow, metabolize, and decay according to environmental conditions. Designing with them requires accommodating biological agency, not imposing rigid parameters. The result is a move away from linear manufacturing toward metabolic cycles. Regenerative intelligence treats these cycles not as constraints but as sources of intelligence—systems that sense, respond, and heal. As climate pressures increase, this form of intelligence becomes a critical design asset.

Mapping Multispecies Realities
If regenerative intelligence asks designers to collaborate with living systems, it also requires a deeper understanding of the infrastructures that shape ecological relations. Feral Atlas, led by a collective of researchers and artists and published by Stanford University Press, offers one of the most comprehensive recent attempts to map these relationships. The project documents “feral” entities—organisms, infrastructures, and hybrid ecologies that have emerged from human systems but exceed human control.

Rather than presenting a stable view of environmental impact, Feral Atlas highlights the micro-politics and material entanglements that shape contemporary ecologies. It underscores how industrial infrastructures, colonial legacies, and global supply chains have produced hybrid ecosystems that cannot be reduced to either nature or technology. This form of mapping expands the scope of regenerative intelligence by revealing where design interventions need to occur and where unintended ecological actors already operate. By articulating how systems behave at multiple scales—from microbial interactions to geopolitical infrastructures—Feral Atlas offers a framework for designers working toward regenerative outcomes. It shows that resilience is not simply a material property; it is produced through entangled relationships.

Building for Resilience, Not Permanence
A critical component of regenerative intelligence is the recognition that resilience emerges from diversity, redundancy, and continuous feedback. It stands in contrast to the architectural pursuit of permanence or the technological pursuit of control. Instead of designing structures that resist change, regenerative systems absorb, redirect, and adapt to it.

Studio Ossidiana’s landscapes shift with tidal rhythms and seasonal variations. Heringer’s buildings invite maintenance as a living process. Feral Atlas reveals the way unintended ecologies form in the gaps of infrastructure. Each of these practices challenges the assumption that stability is the ultimate design goal.
Regenerative intelligence suggests that the environments most capable of withstanding climate uncertainty are the ones that can reorganize themselves. These systems are not simply efficient; they are alive in the sense that they maintain relationships, exchange matter and energy, and renew themselves over time.

For designers, this translates into a new set of priorities: designing for feedback, fostering ecological co-production, and planning for ongoing transformation. The shift is subtle but significant. Instead of asking how design can control environments, the question becomes how design can participate in ecological regeneration. As more practices engage with living systems—through biofabrication, community-embedded construction, feedback-driven landscapes, and multispecies mapping—regenerative intelligence is emerging as a foundational lens for designing the future.

Interfaces now operate as policy. Every toggle, dropdown, and confirmation flow encodes a worldview—sometimes intentionally, often by default. As digital systems scale across finance, civic infrastructure, and everyday communication, interaction design has become an engine of soft governance. Not the legislative kind, but the procedural rules that determine how people coordinate, what actions are possible, and who gets to participate.

This shift isn’t ideological. It’s structural. Distributed networks depend on UI scaffolding to make complexity manageable. Corporate platforms use interface architecture to frame what counts as consent. Civic tools convert municipal processes into clickable workflows that either expand or shrink public agency. Governance is migrating from law to layout. The question is no longer whether design shapes power. It’s how deliberately we’re willing to treat it as a governance layer.

Interfaces as Protocols: The Narrow Bandwidth of Decentralization
Blockchain systems make this visible. Ethereum’s mechanics are abstract, but wallets, dashboards, and explorers function as translation layers that define how the network is experienced. These aren’t neutral windows—they’re governance surfaces.

Designer and researcher Kei Kreutler describes how DAO interfaces shape the mental models of collective ownership. The metaphors—treasuries rendered as financial dashboards, voting as a slider bar, membership as a colored graph—preset expectations about what participation means. When the UI compresses governance into a single action, it narrows the imagination of what governance could be.

Artist and developer Sarah Friend extends this further by designing systems where participation requires sustained care or social negotiation. Her work makes visible how UI mechanics can enforce norms: a button that expires, a screen that requires collaboration, an interaction that demands accountability. The interface becomes a rulebook, not an illustration.

Decentralization often promises openness, but UI constraints determine what “open” actually looks like. A multisig workflow that’s confusing or rigid reshapes who can coordinate. A wallet recovery path that requires technical fluency alters who feels safe participating. Governance, in practice, becomes whatever the interface allows.

Corporate UX and the Quiet Politics of Permission
At the opposite end of the spectrum sit tightly controlled ecosystems. Here, governance emerges not through collective coordination but through UI-defined boundaries. Apple’s privacy settings are a prime example. When the company pushed tracking permissions into explicit pop-ups, mobile advertising economics shifted almost instantly. A modal window became a policy instrument powerful enough to reroute revenue across entire industries. That’s interface as regulation, executed through design rather than legislation. But Apple’s patterns also show the limits of corporate-defined agency. Whether a user selects “allow once,” “allow while using the app,” or “don’t allow,” the interface frames privacy as a series of choices that fit Apple’s architecture—and Apple’s business priorities. What isn’t presented isn’t considered.

Designer and researcher Caroline Sinders has examined how reporting tools and moderation flows on major platforms codify permissible forms of expression. The labels in a dropdown menu define the boundaries of harm. The number of steps in a submission process influences whether people speak up. These structures aren’t decorative; they’re forms of governance embedded in UX. Corporate UX, then, operates as a policy membrane. It decides how much friction users encounter, which decisions require confirmation, and which are silently automated. Each choice carries political weight.

Civic Interfaces and the Architecture of Participation
Cities are increasingly governed through portals, dashboards, and consultation tools. Civic tech aims to make processes legible, but the interface often becomes the primary gatekeeper.

Architect and urbanist Pablo Sendra, alongside sociologist Richard Sennett, has explored how participatory systems fail when they simulate openness without enabling actual influence. The translation of civic processes into interface elements—the comment box, the vote button, the feedback map—can either democratize decision-making or reduce it to symbolic input.

Researcher Chris Speed’s work in digital place-making shows how design choices affect local agency. Interfaces can foreground community narratives, or they can abstract them into datasets. They can amplify local knowledge, or flatten it in the name of efficiency. The UI becomes a mediator between public values and institutional workflows. Net art pioneer Olia Lialina offers another important legacy. Her advocacy for the expressive, empowered “user” stands as a counter-model to contemporary platforms that treat individuals as endpoints. Civic interfaces that draw from this lineage expand agency; those that ignore it risk eroding it. In each case, design decides which publics count.

Designing Power: Toward More Negotiable Systems
When governance moves into the interface layer, design gains new responsibilities. Interaction patterns become policy instruments. Screen flows become behavioral constraints. Defaults become political positions.

The challenge is not to eliminate this power—governance is unavoidable—but to design for systems that remain legible, adjustable, and open to critique. For decentralized networks, that means interfaces that expose rather than conceal complexity, allowing communities to renegotiate structures. For corporate platforms, it means acknowledging that privacy, security, and agency are shaped through design decisions, not legal disclaimers. For civic systems, it means building UI architectures that expand—not compress—public influence. Interfaces are no longer front ends. They are governance engines. And as more decisions migrate into digital systems, designers increasingly shape the rules of engagement. The future of governance will be built in screens, flows, and interaction models. The question is whether we build them with accountability—or simply inherit them as default infrastructure.

Google has introduced Antigravity, a new development environment built around the capabilities of Gemini 3. Rather than functioning as a traditional code assistant, Antigravity positions autonomous agents as active participants in the development process—capable of writing code, running terminals, navigating a browser, and generating verification artifacts within a single interface.

The platform offers two working modes. Editor View mirrors a conventional IDE but adds a persistent agent sidebar for planning and task execution. Manager View expands this into a coordination layer where developers can supervise multiple agents working across different workspaces. The system is designed for structured delegation: users outline a task, agents carry it out in stages, and Antigravity records the process through task lists, screenshots, and browser captures.

While Gemini 3 is the default engine, Antigravity also supports selected third-party and open-source models, signaling Google’s intent to position the environment as a flexible platform rather than a model-locked tool. Antigravity launches in public preview with generous rate limits, giving developers early access while allowing Google to evaluate how autonomous agents perform in real-world conditions. Early feedback will likely influence how agent autonomy evolves.

For now, Antigravity’s impact is practical. It accelerates setup tasks, testing routines, documentation lookup, and iterative coding. Its agent-first structure introduces a more traceable and coordinated approach to AI-assisted development than previous generations of coding assistants.

When Affinity relaunched as a unified, no-charge design application this week, it sent a ripple through the creative-software market. The shift from a paid-license model to a “free-forever” core app—while gating select AI features behind a separate subscription—places the product squarely in competition with Adobe’s subscription ecosystem.

The new Affinity app merges photo editing, illustration, and layout tools into a single workspace, integrating much of the functionality of Affinity Photo, Designer, and Publisher. Designers are already testing whether it can serve as a daily-driver alternative to Photoshop, Illustrator, or InDesign. While many welcome its zero-cost entry point, others remain cautious about how the freemium model—and the reliance on premium AI tools—will sustain long-term development and professional reliability.

By offering a professional design suite at no cost, Affinity is challenging long-held assumptions about creative-tool economics. The move pressures legacy platforms to justify subscription pricing and raises broader questions for designers navigating increasingly AI-augmented workflows: Who controls the toolchain? Who owns the data? And what does creative autonomy mean when the platform itself is free but monetized elsewhere?

Whether Affinity’s strategy sparks a broader shift or simply establishes a loyal open-access community, it underscores a pivotal shift in how creative-technology companies compete—not just on features, but on access, interoperability, and trust.

A team led by researchers at Massachusetts Institute of Technology (MIT) has developed a new algorithmic layer that harnesses generative-AI techniques to design quantum-capable materials — shifting generative art beyond pixels and into atomic architectures.

Traditional generative-AI models in materials science excel at proposing vast libraries of compounds but struggle to home in on the atomic lattice geometries that yield quantum phenomena like spin liquids or flat-band magnetism. The MIT group’s solution, dubbed “SCIGEN,” embeds physics-based constraints into diffusion-model workflows, steering the generation process toward crystal motifs — for example, Kagome or Lieb lattices — that foster exotic magnetic behaviour. 

In practice, the pipeline generated over 10 million candidate structures, sorted them via stability filters and high-fidelity simulations, and ultimately yielded at least two synthesized compounds whose measured magnetic characteristics matched predictions.The result offers a blueprint for bridging generative design, materials science and quantum technology.

MIT’s SCIGEN research marks a genuine inflection point. The algorithm doesn’t just generate images or interface layouts — it generates atomic structures engineered for specific physical behaviours. By embedding structural and physics-based constraints into diffusion-model workflows, the system effectively lets AI propose stable crystal frameworks with target magnetic or quantum properties. MIT researchers suggest that such workflows could soon integrate with autonomous lab systems, enabling an end-to-end pipeline from generative model to synthesized material — a tangible step toward AI-assisted quantum materials engineering.

Industrial designer Ross Lovegrove has teamed with Google DeepMind to chart a new territory where generative-AI meets industrial prototyping. According to a blog post from the research lab, the collaboration involved fine-tuning a text-to-image model (based on Imagen) with a curated dataset of Lovegrove’s sketches and design lexicon — enabling the system to internalise his signature biomorphic forms.

The project selected a chair — a conventional object with defined function but open formal possibilities — as the testbed. Lovegrove’s studio, in concert with DeepMind engineers, developed a design vocabulary and deliberately avoided the word “chair” in prompts to coax novel outputs. From concept output to physical manifestation, the workflow passed through AI generation, material and viewpoint exploration using Google’s Gemini tools, and then to metal 3D printing.

The outcome: a tangible prototype that embodies Lovegrove’s aesthetic, yet extends it through human-machine co-creation. The case indicates a shift in industrial-design workflows — not AI replacing the designer, but augmenting the designer’s generative capacity. As Lovegrove puts it, “For me, the final result transcends the whole debate on design. It shows us that AI can bring something unique and extraordinary to the process.”

For design professionals, this signals that mastering prompt-crafting, semantic modelling of a design language, and physical-manufacturing pipelines are becoming core skills. In operational terms, the experiment maps a workflow: dataset → fine-tuned model → prompt-driven generation → CAD refinement → additive fabrication — a template for future studio-AI collaborations.

Over the past two decades, generative design and digital manufacturing have reshaped how furniture and material systems come into being. What began as a series of experimental prototypes has evolved into a mature design ecosystem—one driven by algorithmic logic, simulation, and robotic fabrication. Form now emerges not from sketch or style but from computation: structures refined through topology optimization, additive manufacturing, and data-driven performance criteria. What was once an industrial process of assembly has become one of emergence.

Generative Intelligence and Material Behavior
At its core, generative design marks a shift in authorship. Instead of drawing fixed forms, designers define parameters—constraints, materials, forces, and goals—and allow algorithms to iterate through thousands of potential solutions. Topology optimization, originally developed for aerospace and automotive engineering, has moved into product and furniture design. It minimizes material while maintaining strength, yielding forms that resemble natural growth patterns: lattices, branches, and bone-like geometries that are mathematically efficient and visually organic.

A landmark example is the AI Chair by Philippe Starck, developed with Kartell and Autodesk. Using Autodesk’s generative design software, Starck set design goals for comfort, stability, and minimal material use. The software produced multiple configurations optimized for those parameters, from which Starck selected and refined the final version. The resulting skeletal form was manufactured through additive processes before being adapted for production. While marketed as co-designed with artificial intelligence, the project more accurately illustrates algorithmic generativity—a designer guiding computation toward structural intelligence.

Joris Laarman’s Bone Chair (2006), created with engineers from BMW’s Rapid Prototyping Division, also embodies this synthesis of algorithmic form and fabrication. Based on Claus Mattheck’s bone-growth simulation, the algorithm removed material where it wasn’t structurally necessary, mimicking biological efficiency. The result became a milestone in computational furniture design and a statement on design as a negotiation between biology, engineering, and aesthetics.

Today, these early experiments have evolved into full-fledged design workflows. Tools such as nTopology and Autodesk Fusion 360 integrate simulation, performance analysis, and generative modeling within a single environment. Designers can test loads, material behavior, and fabrication parameters in near real time, allowing form to evolve with mechanical data. The studio increasingly resembles a hybrid lab—where intuition, code, and material computation co-produce new forms of making.

Printing Structures, Not Surfaces
Additive manufacturing has matured from prototyping tool to production method. For furniture designers, it enables geometries that traditional fabrication cannot achieve, especially when paired with generative modeling. The result is not just new aesthetics but new material logics.

 Hot Wire Extensions, originally developed by Fabio Hendry and Christoph Hauf at the former Studio ilio, exemplifies this shift. The London-based designers introduced a process of heating recycled nylon powder with a custom hot-wire rig to create furniture with a porous, coral-like texture. Rather than concealing the irregularities of additive manufacturing, the designers amplified them—turning digital imperfection into material expression.

By contrast, The New Raw, a Rotterdam-based studio founded by Foteini Setaki and Panos Sakkas, explores large-scale robotic printing with recycled plastics. It's Print Your City! initiative invites citizens to transform plastic waste into urban furniture—benches, planters, and playground structures—via parametric design software. Users customize form and color before fabrication, revealing a direct feedback loop between data, participation, and physical outcome.

Hybrid fabrication pushes this logic further. Zaha Hadid Architects’ Striatus Bridge (2021), designed with ETH Zurich’s Block Research Group and Incremental3D, demonstrates how robotic extrusion and structural modeling can produce complex geometries assembled without adhesives. Though architectural in scale, it illustrates a principle that furniture designers are adopting for modular joints and adaptive assemblies—where variation within constraint replaces repetition as the guiding logic.

Two decades after these early explorations, generative furniture design has entered a new phase. Studios such as Nagami and Gantri are extending these principles into scalable production systems, combining computational modeling with circular materials and distributed manufacturing. Nagami’s Circular Lab (2023–24) advances closed-loop polymer research and AI-driven form optimization, while Gantri continues to scale 3D-printed lighting production through parametric customization and sustainable, plant-based materials. Together, these developments signal a shift from isolated experimentation to system-level transformation—where generative principles are embedded within industrial production and everyday design.

Data as Craft
Digital manufacturing often seems to replace the hand with the machine. Yet across contemporary practice, it has evolved into a new kind of craftsmanship—a dialogue between data and material. The designer’s role shifts from manual execution to curatorial intervention: defining parameters, refining algorithmic outcomes, and interpreting feedback between digital process and physical result.

Patrick Jouin’s Solid C2 Chair (2004) exemplifies the early days of this negotiation. Produced through selective laser sintering (SLS) in collaboration with Materialise, it was among the first pieces of 3D-printed furniture to move beyond the prototype stage. Jouin digitally modeled the chair as part of his Solid Collection, exploring how additive manufacturing could produce seamless, complex geometries impossible to achieve through traditional molding or assembly. The resulting piece—a single, lightweight form with intricate internal structures—demonstrated how digital fabrication could extend, rather than replace, the designer’s craft.

The Growth Chair by Mathias Bengtsson series extends this computational materialism further. Developed through custom software simulating cellular growth, the works are fabricated in composite materials or cast from 3D-printed molds. Bengtsson treats code as a living agent, allowing form to evolve through virtual processes analogous to natural morphogenesis. The resulting structures blur the line between digital automation and artisanal authorship, showing computation as an extension—not an erasure—of craft.

In parallel, Layer Design (2023) is developing AI-assisted seating prototypes generated through cloud-based simulation. These projects suggest how human–machine co-creation is entering mainstream industrial design, translating the craft of computational iteration into adaptable, mass-customizable systems.

Toward Adaptive Fabrication
What connects these projects is not the tools themselves but the paradigm behind them: a move from fixed design to adaptive systems. Generative algorithms, simulation engines, and robotic fabrication all operate through feedback loops, iteratively responding to data to refine structure, performance, and material behavior. This convergence enables a new kind of design ecology in which material distribution, assembly, and even end-of-life scenarios are driven by performance metrics rather than static convention.

At ETH Zurich’s Digital Building Technologies (DBT) group, researchers are advancing these principles at architectural scales, using 3D-printed molds and computational workflows to optimize concrete components for strength, efficiency, and form. Projects such as Smart Slab (2018) and Concrete Choreography (2019) demonstrate how algorithmic design and robotic fabrication can minimize material while maintaining structural integrity. Though developed for buildings and pavilions, these methods point toward near-future applications in furniture and product design—where objects could be locally printed on demand and tuned to specific users, materials, and contexts.

Generative design is not merely producing new aesthetics; it is redefining how designers conceive of making itself. The emphasis shifts from static form-giving to dynamic system-tuning—from the object as an endpoint to design as an evolving process. Data becomes a design material, shaping how density, mass, and performance are distributed across a structure. The furniture that emerges from these workflows is not simply digitally fabricated—it is computationally conceived, the product of continuous negotiation between algorithm, material, and maker.

The global market for generative-design tools could reach US $8.45 billion by 2029, according to a new forecast from The Business Research Company (TBRC). The report projects a compound annual growth rate of about 17 percent, driven by advances in artificial intelligence, cloud computing, and simulation technologies.

Generative design—once a niche within engineering and architecture—is now influencing manufacturing, product development, and creative workflows. The technology uses algorithms to explore thousands of design iterations, optimizing for parameters such as performance, material efficiency, or cost. TBRC notes that adoption is accelerating across industries from automotive and aerospace to construction and consumer products, particularly with the rise of AI-assisted modeling and additive manufacturing.

Other firms offer different market projections, reflecting varying definitions. Straits Research estimates the generative-design market could reach US $11.87 billion by 2033 (CAGR ~16 % from 2025). Meanwhile, Fortune Business Insights projects a smaller figure of US $926 million by 2030 (CAGR ~20.1 % from 2023), reflecting its narrower scope.

For designers, the upward trend points to a deeper shift in authorship and workflow. As algorithms evolve from assistants to collaborators, generative design is becoming less a technical method than a creative infrastructure—one reshaping how form, function, and computation converge across disciplines.

Adobe Inc. has unveiled AI Foundry, a new platform enabling enterprises to build custom generative-AI models tuned to their own brand assets, tone, and creative identity. Announced in October 2025, the initiative extends the company’s Firefly model family into an enterprise-grade framework designed to align with creative workflows across Adobe Creative Cloud and Experience Cloud.

Rather than a self-service platform, AI Foundry operates as a managed collaboration between Adobe and enterprise clients. The company works directly with organizations to develop deeply tuned models using proprietary brand data—ensuring results that reflect specific visual languages and corporate identities. These customized Firefly derivatives can generate brand-consistent images, videos, and 3D assets while maintaining legal and ethical safeguards around data provenance and intellectual property.

For creative departments, the implications are substantial. Design teams now have access to bespoke generative systems that are tuned to their brand’s visual vocabulary, narrative tone, and campaign parameters. Rather than generic tools, this offering positions AI Foundry as an extension of the creative stack—one that collaborates with human teams and Adobe experts alike, and is built to scale content production while preserving brand integrity and creative oversight.

The launch signals a broader shift in creative workflows: design teams are increasingly acting as orchestrators, leveraging brand-specific AI models rather than generic tools. With AI Foundry, Adobe is betting that enterprise creativity will evolve around customized AI systems built to each organization’s unique identity.

The boundaries between art, design, media, and technology are dissolving. Once distinct disciplines now operate within an expanded field where digital systems and material environments merge into new forms of hybrid creativity. The result is not simply a digital transformation but a shift in how creative practice itself is conceived—distributed across software and sensors, data and gesture, virtual and physical space.

From interactive installations and robotic sculpture to generative architecture and responsive environments, today’s design-tech landscape is defined by convergence. Artists and designers are no longer working with technology—they’re working through it, using computation and connectivity as the raw material of expression.

Hybrid Interfaces
The most compelling examples of this continuum treat technology not as a tool but as a collaborator. Studio Drift’s Fragile Future series, for instance, merges dandelion seeds and LED circuits into delicate light sculptures that fuse natural fragility with technological precision. The work makes tangible the interdependence of organic and artificial systems—an aesthetic logic echoed in their recent drone-based performances like Franchise Freedom, where hundreds of autonomous drones mimic the collective behavior of starlings in flight.

Similarly, Random International’s Rain Room (2012–ongoing) transforms environmental control into an embodied interface. Using motion sensors and real-time data processing, the installation allows visitors to walk through falling rain without getting wet, collapsing the distinction between environmental experience and algorithmic control. The body becomes a responsive node within a cyber-physical system.

Such works signal a broader trend: interaction as material. Rather than static displays, creative systems now respond dynamically to movement, sound, or data. This is evident in the installations of Refik Anadol, whose Machine Hallucinations series translates massive datasets—ranging from architectural archives to satellite imagery—into immersive, AI-generated visual environments. These works are not representations but processes, blending computational abstraction with sensory immersion.

The Living Circuit
In parallel, designers and technologists are exploring how biological and synthetic systems might inform one another. MIT’s Mediated Matter Group, led by Neri Oxman, pioneered this approach through projects like Aguahoja (2018), a series of biodegradable structures fabricated from chitin, cellulose, and pectin using robotic extrusion. The result is a material ecology that repositions design as a living interface between nature and computation.

This bio-digital approach extends to projects like Hylozoic Ground by Philip Beesley, where sensor-embedded structures appear to “breathe” in response to human presence. Here, architecture behaves as an organism—a responsive ecosystem driven by feedback loops between machine intelligence and environmental stimuli.

Across these works, technology becomes less about efficiency and more about emergence. Designers operate as system choreographers, setting up conditions for interaction and adaptation rather than dictating fixed outcomes. The resulting artefacts—whether soft robotics, kinetic architecture, or responsive textiles—embody a new kind of intelligence, one that is distributed, embodied, and situational.

Immersive Systems
Media artists have long experimented with immersive environments, but recent advances in sensing, real-time rendering, and spatial computing are expanding the field’s vocabulary. Tokyo-based collective teamLab has become a global reference point for this mode of practice. Their installations—such as Borderless in Tokyo and Planets in Osaka—dissolve the physical boundaries between viewer and artwork. Motion tracking, projection mapping, and multi-sensory cues generate a seamless field of interaction, where digital flora respond to touch and presence.

Beyond spectacle, immersive systems are being used as instruments for inquiry. Marshmallow Laser Feast’s We Live in an Ocean of Air (2018) visualizes the exchange of oxygen and carbon dioxide between humans and trees using real-time data and VR. By translating invisible ecological processes into perceptual experiences, the work reframes immersion as a medium for environmental awareness rather than escape.

Meanwhile, designers like Es Devlin are using AI-driven projection and choreography to stage performances that integrate audience movement and collective data. Her Memory Palace (2019) installation at Pitzhanger Gallery transformed historical narratives into spatial storytelling, blending physical architecture and digital mapping. These projects foreground narrative as a spatial, networked experience—an architecture of information as much as of form.

Designing the Continuum
What ties these practices together is a shift from static object-making toward systems thinking. The creative act is no longer confined to producing artifacts but to designing the conditions through which matter, data, and behavior interact. The continuum between digital and physical is now an operational space—where feedback, automation, and embodied experience converge.

This perspective is increasingly visible in architecture and product design. Firms like MAD Architects, Zaha Hadid Architects’ CODE group, and MX3D are developing adaptive systems that combine parametric modeling, robotic fabrication, and environmental sensing. MX3D’s 3D-printed steel bridge in Amsterdam, equipped with sensors and data visualization tools, stands as a literal manifestation of cyber-physical infrastructure—a structure that monitors its own behavior while serving as public space. In industrial design, practitioners such as Ross Lovegrove and Formafantasma explore material intelligence through digital modeling and sustainable processes. Their works translate the logic of computation into tangible form, blurring the line between algorithmic precision and artisanal craft.

As boundaries dissolve, the creative process becomes less about authorship and more about orchestration. The designer’s role shifts toward that of mediator—navigating between hardware and software, code and culture, human and non-human systems. The challenge is not simply to integrate technology but to question how it redefines creativity, ethics, and agency within interconnected environments.

Toward Cyber-Physical Practice
The convergence of art, design, and technology is not a passing phase but a structural evolution of creative culture. As AI, robotics, and sensing technologies permeate material and spatial practice, the question is no longer whether technology belongs to design, but how design can responsibly shape hybrid worlds.
Cyber-physical creativity points toward futures where the boundaries between virtual and tangible dissolve, where data behaves like matter, and where imagination operates as both biological and computational intelligence. From drones that mimic nature to structures that sense and respond, the design-tech continuum signals a cultural shift: creativity as a dynamic interface—alive, adaptive, and shared between human and machine.

Synthetic landscapes rethink the relationship between design and ecology. At the Southern California Institute of Architecture (SCI-Arc), the term defines a field where the artificial and the natural can no longer be separated. The program describes the built environment as “always continuous with what we have found,” positioning design as part of a living, evolving system rather than something imposed upon it.

This approach departs from traditional landscape design focused on preservation or stewardship. Instead, it examines how architecture, computation, and environmental systems intersect to generate new forms of coexistence and beauty. Design operates within networks of human and non-human agents, linking the biological and the technological to imagine how future environments might take shape. From speculative ecosystems that respond to pollution to living architectures that grow and adapt, these projects suggest that ecology itself has become a designed medium—revealing how entangled our built and natural worlds have always been.

Speculative Ecosystems
Artist-designer Pınar Yoldas frames the polluted oceans as a “plastic soup” and asks: if life were to begin now, in that synthetic milieu, what organisms might evolve? Her installation, An Ecosystem of Excess, imagines life forms capable of metabolizing plastic waste, offering a speculative biology of the Anthropocene rather than a “restored nature.”

By treating waste and human-technological detritus as ecological media, Yoldas reorients the landscape—not as something to preserve, but as something already irreversibly altered and now to be designed. Her work exposes how synthetic landscapes can emerge through cultural-ecological reflexivity: design that interrogates our ecological turning point rather than simply mitigates it.

Bio-Hybrid Built Form
On the more technical end of the spectrum, the EU-funded Flora Robotica project explores symbiotic relationships between plants and robots to prototype forms of living architecture. Supported under the Horizon 2020 program (Grant No. 640959), the research team developed systems where robotic structures braid soft scaffolds, guide plant growth, and sense environmental change. As the plants mature, they overtake and integrate these braided supports, forming hybrid constructions that are both biological and artificial.

Flora Robotica’s publications describe this as an “architectural system combining living natural plants and distributed robots,” in which “plants and robots work together to create a living organism that is inhabited by human beings.” The project investigates how robotic control, plant tropisms, and environmental sensing can generate architectural forms through continuous growth rather than assembly.

The result is a model of design based on process and co-evolution. Architecture becomes something that grows, adapts, and repairs itself—guided not only by human intention but by the behavior of living systems. Flora Robotica situates architecture within a dynamic ecology of sensors, materials, and biological intelligence, suggesting how the built environment might one day sustain itself through collaboration between machines and living matter.

Generative Ecosystems
Beyond fixed installations and architecture, another trajectory of synthetic landscapes emerges through generative systems that integrate living, algorithmic, and material flows. The Synthetic Ecologies Compendium—organized by the Serpentine—centres on “fermentation, invisible scales of life and the intersections of biology, culture and computation.”

More recently, the research project Simulacra Naturae combines neural activity, agent-based simulation, and living plants to create a multisensory ecosystem where biological processes and generative code evolve together. These systems shift the design question from formal composition toward emergent dynamics—designing for process, thresholds, and interaction rather than static objects. The synthetic landscape here is less a site than a looping system of feedback among living, artificial, and computational agents.

Design as Ecology
Across these projects, ecology becomes a design interface—adaptive, coded, and alive. In Yoldas’s An Ecosystem of Excess, plastic-polluted oceans become a habitat for speculative organisms that metabolize synthetic waste. In Flora Robotica, plants and robots braid scaffold structures and guide each other’s growth toward architectural form. And in Simulacra Naturae, neural activity, agent-based simulation, living plants, and generative systems intertwine to create evolving environments of material, biological, and computational feedback. Each reframes “landscape” not as a static backdrop but as an ever-shifting system of interaction.

In design studios and research labs, the shift is tangible. Climate and soil data flow into texture maps. Neural networks train on sap flow and plant electrophysiology. Bioreactors and hydroponic systems feed fabrication workflows where code directs photosynthesis with the same precision once applied to CNC milling. Architecture and computation move beyond representation—they become ecological participants.

As synthetic landscapes proliferate, authorship disperses. The designer’s hand gives way to calibration: tuning parameters, guiding microbial metabolisms, steering robotic motion, adjusting generative algorithms. What emerges is not an imitation of nature, nor its conquest, but a mutual composition—an ecology written with code, mycelium, and design intelligence.

OpenAI’s newly launched ChatGPT Atlas merges web browsing with conversational AI. Its most distinctive—and controversial—feature is “Memories,” a function that allows the browser to retain contextual information about a user’s activity across sites. The goal is to create continuity: Atlas can recall prior searches, revisit ideas, or suggest relevant actions based on past behavior. But this same continuity is sparking concern among technologists and creatives who see in it a potential for overreach.

Data Retention Meets Personalization
OpenAI states that browsing content isn’t used to train models by default and that users can view, edit, or delete their stored data. Yet, “memory” transforms a transient browsing act into a persistent behavioral record. For designers, researchers, and creative professionals whose workflow depends on iterative exploration, this could blur lines between assistance and surveillance. The convenience of context-aware AI comes with the trade-off of deeper data visibility—how much the browser knows about your habits, projects, and even creative process.

Trust as Interface
The question isn’t just technical—it’s relational. Atlas redefines the browser from a neutral interface to an active participant in knowledge work. The tension lies in trust: can users rely on a system that both assists and archives? As OpenAI pushes toward agentic AI—browsers that act rather than observe—the debate over memory, privacy, and agency will shape the future of creative technology. For now, Atlas offers a preview of a more intimate web, one that remembers as much as it serves.