When Art Computes: New Media on the Edge of Expression

In 2018, Portrait of Edmond de Belamy by the Paris-based collective Obvious became the first AI-generated artwork to sell at a major auction, fetching $432,500 (USD). While the work’s GAN-generated aesthetic received mixed critiques, the sale signaled a broader transformation: computation is no longer just a tool for making art — it is now a collaborator. Artists today are not merely adopting technology but embedding it into the practice itself.

The new frontier includes works like TeamLab’s immersive installations, where fluid digital projections interact with physical space, drawing audiences into dynamic, co-created environments. The Japanese collective’s Borderless exhibition at Tokyo’s Mori Building Digital Art Museum, launched in 2018, registered over 2.3 million visitors in its first year. TeamLab’s approach extends beyond visuals; by using sensors, projectors and algorithms, their pieces react to human movement, reframing spectators as participants.

But the shift is not confined to large-scale installations. Individual artists are also pioneering deeply technical approaches. Harshit Agrawal, a Bangalore-based new media artist, blends machine learning with traditional Indian art forms to explore themes of identity and culture. His series EXO-Stential AI (2019–2020) leverages GANs trained on handcrafted patterns, merging ancient aesthetics with code to pose questions about originality in an increasingly synthetic age.

The rapid development of generative AI, particularly since the release of OpenAI’s GPT-3 in 2020, has further blurred the lines between artist and algorithm. However, the integration of these tools invites questions about authorship, labour and ethics. Dr. Joanna Zylinska, a scholar of new media at King’s College London, has written extensively on what she calls "nonhuman creativity": the idea that technological systems produce outputs that cannot be entirely attributed to human intention. According to Zylinska, "The critical question is not whether machines can make art, but how their outputs reflect the systems of power and bias they are trained on."

Interactive installations, meanwhile, are pushing interactivity itself into conceptual territory. In Nairobi, the 2022 project Data Bodies by artist and technologist Jepchumba created a multiphase installation that required visitor data for entry. Participants submitted anonymised metadata — from browser histories to location pings — to generate personalised visual projections, which represented their digital lives as ephemeral, shifting patterns. "The intent," Jepchumba noted at an artist talk, "was to make the invisible visible, forcing participants to confront how their data is commodified and abstracted."

Such works are actively rewriting the grammar of art. Whereas traditional mediums like oil or marble demanded permanence, digital and interactive media emphasize the fleeting. Even the infrastructure of such works is unstable: server migrations, software deprecations, and hardware failures threaten their longevity. Yet artists and technologists alike are developing responses. Rhizome’s Emulation-as-a-Service platform, launched in 2013, offers one solution by preserving web-based artworks through virtualised environments, keeping works like My Boyfriend Came Back from the War (1996) by Olia Lialina accessible, decades past their original technical contexts.

Institutions, meanwhile, are grappling with how to exhibit and archive digital art responsibly. The Whitney Museum of American Art’s acquisition of Cloud Cities: Networked Art in the Era of Social Media (2019) included not only the works themselves but extensive metadata, source code and mirror servers to ensure their reproducibility. Such moves reflect a growing realization that the medium itself carries meaning — a JPEG saved as a blockchain-hosted NFT is not the same as one on an artist’s private server, regardless of identical pixels.

The commercial sector is not idle, although it often prioritizes spectacle over critical engagement. In 2021, Refik Anadol’s Machine Hallucinations: Space became one of the most commercially successful NFT-driven installations, auctioned for $4.4 million (USD) via Sotheby’s. While critics such as Anil Dash, co-creator of the original NFT standard, have pointed out the environmental cost and speculative excess of blockchain-backed digital art, it remains part of the technological infiltration into creative economies.

The intersection of art and technology remains a contested space. Some argue it suffers from a techno-utopian rhetoric that prioritizes novelty over substance. Others, like curator Nora Khan, see it as an expansion of the artistic toolkit, empowering creators to engage directly with systems of surveillance, datafication and algorithmic control. Khan’s 2023 essay Art in the Age of Computation (e-flux Journal) outlines how artists can "turn the tools of power into instruments of critique."

What the field offers, ultimately, is not just new media but new paradigms of interaction. Visitors to Studio Olafur Eliasson’s Your Digital Shadow (2023), exhibited at the Pinakothek der Moderne in Munich, found themselves projected as abstracted avatars — digital doubles that mimicked their movements but distorted them, introducing deliberate glitches and delays. The work posed a simple question: who owns your shadow?

This confrontation between the self, the tool, and the systems surrounding them defines much of contemporary digital practice. It suggests that as technology embeds itself into artistic production, the artworld itself becomes a site of negotiation between machines, humans and the data that entangles them. The boundaries are unlikely to settle soon.