The World’s First AI Art Museum Has a Strange Way of Honoring the Rainforest

It’s no surprise that the world’s first museum of AI art isn’t a museum in the traditional sense. Dataland—co-founded by husband-and-wife artist duo Refik Anadol and Efsun Erkiliç—is nestled in downtown Los Angeles’ museum mile, where massive open-source models, along with visitors’ real-time biometric data, project maximalist AI-generated installations across its galleries.

It’s also no surprise that those installations rely on the breathless tropes of immersive art experiences: 360° visuals; cinematic soundscapes; engagement of other senses such as smell and taste. But it was at least theoretically possible that a new art museum, designed around a relatively new artistic medium and by one of the highest profile artists working in that medium, might imagine new aesthetic or institutional possibilities for an art world mired in old paradigms.

Unfortunately, Dataland’s inaugural exhibition, “Machine Dreams: Rainforest,” conjures nothing more than turbo-charged versions of familiar spectacles. The very name “Dataland” is a clue: it sounds like an amusement park, and aspects of the museum experience certainly recall one. In the “Discovery Portal,” the darkened antechamber preceding the main galleries, each visitor scans their admission ticket’s QR code on a sealed black box. Then, its panels slide open like a weapons cache in a futuristic spy movie.

Inside the box, there are two objects: a wristwatch that tracks its wearer’s biometric data, such as heart rate and skin temperature; and a neckpiece that intermittently releases scents created by L’Oréal Luxe. While many theme park rides and immersive art experiences incorporate scents these days, the twist here is that the AI model releases fragrances personalized to you based on the data collected from your watch. Yet there’s no discernible rhyme or reason as to why or when visitors are sporadically hit with specific mossy or floral scents.

This sort of gimmick could have been something with substance and bite if Dataland was willing to play with the implications of AI surveillance in provocative ways. Instead, digital wall texts tout the artworks’ responsiveness to visitor data, even as the connection between the work and said data remains either too tenuous or too obvious. The concluding “Sanctuary” gallery, for instance, produces a collective “living portrait” based on the cumulative data of every visitor standing in the room. But its dense wall of graphs appear indecipherable to a layperson, and its morphing blobs of color closely resemble Anadol’s 2022 AI-generated remix of MoMA’s collection, Unsupervised—which is to say, the blobs could represent almost anything. Critics panned Unsupervised as a highbrow screensaver or lava lamp, and though AI has developed a lot in the intervening years, Anadol’s approach has not.

Critical opinion has hardly stopped Anadol’s studio. Numerous perceptive, even sympathetic critiques have been made about his practice, and Dataland takes none of them to heart. It contains more of the same high-powered, low-substance artwork—only now, at a greater scale. The for-profit museum makes clear, if there were any doubts, that Anadol’s business talents far surpass his artistic ones. If Dataland portends a new cultural paradigm, it’s one in which an artist’s studio becomes so resource hungry that it morphs into a tech start up—one fueled less by Venture Capital than by collaborations with bigger companies, like Google, Siemens, and NVIDIA.

AT THE HEART OF the museum, the “Data Pavilion” gallery is a long, tall room designed a bit like a dance hall, with mirrored columns and ceilings refracting sweeping psychedelic nature imagery projected across the room’s vast surfaces. At the press preview remarks, Erkiliç cited the Light and Space movement as an art historical precedent influencing the museum—but this is Light and Space on steroids and made woefully literal. One moment, thousands of luminous globules—molten oranges and reds; deep greens and purples—roil around the room like waves crashing ashore. In another, an enormous grid of oversaturated flower images fragments into pixelated shards of color.

These effects are neat if you enjoy digital pyrotechnics, and the atmosphere is made more dramatic by the orchestral swells and organ drones that serve as their score. Perhaps there’s even an allegory here about how AI technologies produce a restless, shapeshifting world. But it’s hard to see how the artwork “demystif[ies] the data” in the way Erkiliç claims; if it does anything, its phantasmagoric outputs re-mystify the training data, converting over a billion realistic nature images into kaleidoscopic nature vibes.

Digital artists have explored the seams between their synthetic medium and the natural world for half a century at least. For his 1980s “Growth” series, for example, Yoichiro Kawaguchi created evolving, computer-generated, 3D biomorphic forms based entirely on mathematical rules, raising questions about whether algorithms function at all like virtual DNA. More recently, the beautiful yet polluted digital waterscapes in Marina Zurkow’s 2025 exhibition at the Whitney Museum of American Art, “Parting Worlds,” used knowing humor to portray our species’ uncomfortable intimacy with its own waste: in one scene, two people in biohazard suits stroll past what might otherwise seem an idyllic lakeside picnic spot. But Dataland’s nature imagery lacks conceptual purpose beyond showing off the model’s computational power.

The ideas are even thinner in the “Infinity Room,” the exhibition’s other centerpiece. A short text introduces Ruwe Pinu, a “glass hummingbird,” forged “when circuits met chlorophyll,” who “was born to guard the rainforest’s final memory”—as if priming us to mourn the very ecosystem AI is poised to harm. The eight minute 360° video then swerves through scalar and perspectival shifts like an amusement park ride. First-person scenes in bird’s-eye-view float the viewer through floral vortexes and the neurological pathways inside Ruwe Pinu’s cranium. Cyberpunk interludes occur within Matrix-esque circuitry, and lush, third-person rainforest panoramas see our bird friend hovering beside a blossoming flower or the rainforest’s “wisdom tree.”

The film climaxes with a discordant climate parable: the wisdom tree bursts into flames, then morphs into a staticky swarm of dispersing birds. The orchestral music comes to a full stop, replaced by spare bird song. A text appears in the night sky: “At the deepest root of the wisdom tree’s memory is a single song. In 1987, the last Kaua’i ‘Ö’ õ bird sang to a partner that no longer existed. He left pauses in the dark for her to answer. We hold the silence for him.”

The sentiment feels jarring because there is nothing silent or contemplative about the Dataland experience. Everything in the museum pulses like fast-twitch muscle fiber, jittery and kinetic. In multiple galleries, the swirling, multi-surface imagery even induces vertigo in the viewer, similar to the perceptual disorientation you feel when stuck in traffic and the cars going in the opposite direction make it seem as though you’re moving backwards. Amid it all, “We hold the silence for him” comes off as a compensatory cliché incapable of recognizing itself as such.

IF ART HISTORIANS—human or otherwise—still exist in the future, the digital art of the early/mid-2020s will appear telling in relation to whatever comes next. Our decade will either be remembered as just a phase, the time when artistic uses of new media such as NFTs and AI didn’t live up to the revolutionary hype. Or else, it will be remembered as the time when few people were prepared for the cultural transformations about to occur. In the meantime, here in the uncertain present, Dataland’s bizarre Ruwe Pinu extinction fantasia betrays fears about humanity’s potential demise in the face of rapid technological progress, while also embracing the mindset accelerating us toward those doomsday fears.

The past decade of AI development has brought about prominent discussions over the technology’s existential risks. In these hypothetical dystopian futures—in which machines concoct creative ways to enslave or exterminate our species—2020s human culture faces a similar fate to Ruwe Pinu: in which we sing sad, semi-aware songs about our impending obsolescence. The difference, which the video overlooks, is that, unlike the bird, we’re the engineers of our own undoing, and our songs tend to sound shrill rather than poignant.

“Machine Dreams: Rainforest” manifests unease with its own loud role in these trajectories, even as it feigns conscientiousness. The museum’s promotional statements take pains to emphasize Anadol’s and Erkiliç’s ethical approach to AI use: training data gathered with consent; developing models from scratch; minimizing environmental impact (they claim, without showing the math, that each museum visitor consumes about as much energy as it takes to charge one cell phone). These would be laudable ethical designs if Dataland’s sensationalist aesthetics didn’t point in a different, more profit-minded direction. The discrepancy makes the ethical claims look like calculated PR moves to head off criticisms commonly leveled against a medium about which the public feels increasingly skeptical.

Anadol has often described his AI artworks as “machine dreams.” But that metaphor obscures the all-too-human hopes and fears evident in Dataland’s outputs. The museum’s inaugural exhibition wants to have things both ways—a full-speed-ahead future that also mourns what’s being left behind, as if the murderer were also delivering the eulogy at the funeral—and ends up with neither. Its awkward compromises beget tepid innovations upon stale cultural products: an immersive experience with negligible responsiveness; a smart museum gift shop whose offerings are personalized to each visitor based on the data collected (with consent!) about them. “Can an artwork feel us back?” asked Anadol during the press preview. Dataland’s answer is yes—and it would like to use its newfound capacities to sell you a custom t-shirt.