Introduction

On the morning of June 20, 2026, a door opened on Grand Avenue in downtown Los Angeles that had never existed in the history of museums. Visitors walked into a 25,000-square-foot space inside The Grand LA — a $1 billion mixed-use development designed by Frank Gehry, sitting across the street from the Walt Disney Concert Hall — and entered what its founders describe as the world's first Museum of AI Arts. They call it Dataland. Inside, there are no paintings on walls, no sculptures on plinths, no rooms of artefacts behind glass. There are five immersive gallery environments where the air carries the scent of a tropical forest, where a 200-channel sound system wraps around the visitor's body, where kaleidoscopic projections respond to the people moving through them, and where the art is not finished. It is alive, continuously generated by artificial intelligence systems trained on millions of data points from sixteen rainforests across the globe, processing real-time visitor responses, and producing outputs that have never existed before and will never exist again in exactly that form.

The technology driving the experience is a collaboration between Refik Anadol — the Turkish-American media artist who spent a decade installing data-driven projections on some of the most famous buildings in the world — and Google, which serves as the technology and creative collaborator powering Dataland's infrastructure through Google Cloud. The inaugural exhibition, "Machine Dreams: Rainforest," runs until January 31, 2027. The museum itself is permanent. And the questions it raises — about what art is, who makes it, what machines can feel, and where the boundary between creative technology and genuine human expression sits — are not going to be settled any time soon.

What Dataland Actually Looks Like Inside

The experience of Dataland begins before a visitor reaches the galleries. At the Discovery Portal — the museum's entry point — each person receives wearable technology that will track their physical and emotional responses as they move through the space. That data feeds back into the generative systems running the exhibitions, which means the art they encounter is, in a meaningful sense, co-created by their presence. No two visits produce identical outputs. No single visitor experiences what the person before them experienced.

The first major gallery space is a large mirrored room where audiovisual elements fill the environment with shifting light and generated imagery. Anadol described the design intention to The Art Newspaper with characteristic precision: "For 5,000 years, we looked at artworks and we felt something, right? Now, my challenge question is: 'Can the artwork feel us back?'" The Infinity Room — an evolved version of one of the studio's most exhibited installations, which has been experienced by more than 10 million visitors in over 35 cities worldwide — returns at Dataland in its most technically advanced iteration, incorporating AI-generated scent alongside light, sound, and Anadol's Large Nature Model to produce what the studio describes as a fully responsive, living environment rather than a static installation.

The Data Pavilion at the centre of the museum is perhaps the most technically extraordinary space. The 200-channel sound system is not playing recorded audio. It is generating a continuous soundscape from the same AI systems that are producing the visual environment, creating a spatial audio experience that moves and responds to the visitors within it. The symphonic music commissioned from the Los Angeles Philharmonic — a partnership that traces back to Anadol's 2018 projection onto the exterior of the Walt Disney Concert Hall, which is literally across the street from where Dataland now sits permanently — is woven into AI-generated sound in real time. Species imagery from the Large Nature Model's training dataset of ecological data is translated into visual compositions that shift as the music shifts. The scent changes as the visual environment changes. In one section, visitors encounter edible chocolate sculptures modelled on rainforest species — a decision that moves the artwork into the body rather than stopping at the eye or the ear.

The Sanctuary, the museum's deepest gallery, aggregates the biometric and emotional data visitors have shared across their journey and uses it to generate a unique data sculpture in real time — a collective portrait of everyone who has passed through that day, rendered in light and sound. It is the one space where the museum's claims about AI as a genuinely co-creative medium become most concrete: the work that exists in that room at any given moment could not have been made without the specific people who are present, and it will cease to exist when they leave.

The Technology Running Underneath

The scale of the computational infrastructure required to run Dataland in real time is not visible to visitors, but understanding it is essential to evaluating what kind of achievement the museum actually represents.

The Large Nature Model, which Refik Anadol Studio has been developing for several years in collaboration with Google, is a foundational AI trained on data collected first-hand from sixteen rainforests across the globe, combined with extensive datasets from institutional partners including the Smithsonian National Museum of Natural History, the Cornell Lab of Ornithology, Getty, iNaturalist, and the Natural History Museum in London. The choice to build a purpose-trained model on nature data rather than general-purpose training sets was a deliberate ethical decision. Anadol described the reasoning to NPR directly: the moment it became clear that the next generation of AI models would be trained primarily on human-generated content, the studio chose a different path. "The best way to achieve responsible curation is to build our own models and be radically transparent about where our data comes from." Every dataset used in the Large Nature Model is documented and attributed.

Google Cloud provides the infrastructure that allows the Large Nature Model to run in real time across Dataland's five gallery environments simultaneously. Gemini, Google's most capable AI system, is integrated into the generative systems producing the audiovisual outputs — the AI that is, in effect, creating the art that visitors are experiencing. A dedicated 10,000-square-foot space within the museum houses the equipment supporting the model and the computational systems behind the experience. That space is not presented as a utility room to be ignored. It is part of the museum's argument: the machine doing the work is as much a part of the artwork as the light on the wall.

The partnership between Anadol and Google stretches back a decade. Their first collaboration was in 2016, when Anadol joined Google's Artists and Machine Intelligence programme as part of its inaugural cohort. They projected the LA Philharmonic's archives onto Walt Disney Concert Hall together in 2018. They visualised Google Quantum AI data in 2020. In 2025, they commissioned the large-scale installation "Machine Dreams: Biophilia" for Google's Mountain View campus, where an early version of the Large Nature Model generated a continuously shifting digital landscape from regional ecosystem data. Dataland is the permanent form that ten years of that collaboration has finally taken.

The Debate That Opened With the Museum

Dataland's arrival has not been universally celebrated, and the critique that has followed it is not trivial or simply reactionary. It cuts to a question that has divided artists, critics, and philosophers since AI-generated imagery became widely accessible, and that Dataland makes impossible to avoid because it places the question in a cultural institution designed to treat the output of that process as art deserving of serious consideration.

Thomas Brummett, a digital artist whose work sits in the permanent collections of the Museum of Fine Arts in Houston, the Philadelphia Museum of Art, and the Museu de Arte Moderna in Rio de Janeiro, stated his objection plainly on social media after the opening: "Let's build a museum based on instructions people give to AI and call it art. It's not and it never will be. At best, it's just second-rate entertainment." His objection is the articulate version of a widely shared intuition: that art requires intention in a specific human sense, that the creativity which produces meaningful work is inseparable from the consciousness that experiences the world and responds to it, and that a system generating outputs by statistical prediction across training data is performing a fundamentally different process from what human artists have always performed.

Barry Threw, executive and artistic director of the Gray Area Foundation in San Francisco, offered a different and arguably more useful framing in his comment to NPR: Dataland is compelling because it renders "complex data as experience." That framing sidesteps the question of whether the outputs constitute art in the highest sense and focuses instead on what the museum demonstrably does accomplish: it makes information tangible in ways that conventional data visualisation cannot achieve, it creates conditions for genuine aesthetic experience, and it uses technology to produce encounters with the natural world — specifically, with the scale and complexity of rainforest ecosystems — that no conventional museum exhibition could deliver.

The Art Newspaper's reviewer captured the specific quality of the experience in terms that resist easy categorisation: Dataland is "part science experiment, part deeply reverent art museum and part immersive theme park." Those three categories are not a contradiction. They are an accurate description of a new kind of institution that does not map cleanly onto any existing model, and the discomfort of that lack of precedent is precisely what makes the debate about whether it constitutes art both genuine and somewhat beside the point. People who walk through the Sanctuary and watch a real-time data sculpture form from the aggregated emotional responses of everyone present that day are having an experience that produces real responses — wonder, unease, recognition, the particular emotion that arises when something vast is suddenly made intimate. Whether the system producing that experience deserves to be called an artist is a different question from whether the experience deserves to exist.

Anadol has not retreated from the questions. His stated position, articulated across multiple interviews surrounding the opening, is that the Dataland project requires radical transparency as its ethical foundation: full disclosure of every dataset, every institutional partnership, every environmental commitment. The museum operates on renewable energy. Its datasets are attributed. Its AI systems are documented. His statement to the LA Times — "AI means possibilities. And possibilities come with responsibilities. We have to disclose exactly where our data comes from" — is not a defensive response to criticism. It is a proposed standard for the entire field.

What Dataland Signals About Where Creativity Is Heading

The opening of Dataland is a cultural event of a specific and unusual kind. It is not the first time AI has been exhibited in a museum context — the ZKM Center for Art and Media in Karlsruhe has been doing this for decades, and MoMA and the Design Museum in London have both presented AI-assisted work in significant exhibitions. It is the first time a permanent institution has been built specifically to house AI-generated art as its sole purpose, at the scale of a major contemporary art museum, in a building designed by one of the world's most celebrated architects, in direct proximity to some of the most established cultural institutions in Los Angeles.

That context is not incidental. The Grand Avenue Cultural District, where Dataland now sits permanently, includes the Walt Disney Concert Hall, The Broad contemporary art museum, and the Museum of Contemporary Art Los Angeles. Dataland has been placed among them, not on the periphery. Its partnership with the Los Angeles Philharmonic integrates it into the established cultural ecosystem of a city that has become, in the past decade, the most important centre for contemporary art in the United States. Its Google Arts and Culture artist residency — which will provide four artists with $25,000 grants and direct access to advanced Google Cloud tools and machine learning models, with work shown at Dataland and on the Google Arts and Culture website — extends the institution's influence into the next generation of digital creators before it has been open a week.

What all of this signals, taken together, is that the question of whether AI-generated experiences constitute genuine art is no longer primarily a theoretical debate being conducted in academic journals and art reviews. It is being answered institutionally, in concrete terms, in a 25,000-square-foot building on Grand Avenue that opened to paying visitors on June 20, 2026, at $49 per ticket, with a partnership structure involving Google, the Smithsonian, the Los Angeles Philharmonic, Cornell Lab of Ornithology, and a decade of collaboration between one of the world's most inventive data artists and one of the world's most powerful technology companies. The institutions have already decided. The question now is what the public decides — whether the experience of standing in a room that is generating itself in response to your presence constitutes genuine aesthetic encounter or elaborate spectacle.

Conclusion

Dataland is, as its founding artist insists, a living question more than an answer. The building it occupies was left as a concrete shell by Frank Gehry — a space for free imagination, as Anadol described it in conversation with curator Hans Ulrich Obrist. What has been built inside that shell is an institution that cannot be adequately understood through any single existing category: not museum, not theme park, not laboratory, not theatre, though it contains elements of all of them. What it is, most precisely, is a permanent space for the encounter between human presence and machine intelligence in the specific register of aesthetic experience — and it exists in Los Angeles, in a Frank Gehry building, next to a concert hall, in the middle of a cultural district, because its founders believe that encounter deserves exactly that institutional weight.

Whether the outputs of the Large Nature Model and Gemini working together through a 200-channel sound system constitute art in the fullest sense is a question that Dataland does not resolve. What it does, with considerable ambition and considerable technical achievement, is insist that the question be taken seriously. The 10 million people who have passed through an Infinity Room installation in 35 cities did not need to resolve the theoretical debate to know they had experienced something. The visitors stepping off the elevator into The Grand LA on a warm California day and finding themselves suddenly inside an Amazon rainforest — surrounded by light and scent and generative sound that is reading their responses and incorporating them into something continuous and new — will not need to resolve it either. The boundary between technology and culture that Dataland sits on has not been eliminated. It has been given a building. And the building is open.