Curator Hans Ulrich Obrist on a Dynamic ‘Year of AI’ at Serpentine Galleries in London

As artistic director of the Serpentine Galleries in London, Hans Ulrich Obrist has championed all kinds of future-facing art. This year, Obrist—whom the New Yorker once called “the curator who never sleeps”—took that one step further by declaring a “Year of AI” at the lauded institution.

Since joining the gallery’s leadership team in 2006, Obrist has worked to turn the Serpentine into a sort of laboratory. Notable experimental projects and exhibitions he has brought to the institution include Cécile B. Evans’s AGNES, commissioned for the Serpentine’s website in 2014; Ian Cheng’s “sentient artwork” BOB (Bag of Beliefs) and Emissaries, a “video game that plays itself,” in 2018; and NEW FICTION, for which KAWS exhibited artworks at Serpentine as well as within the digital world of Fortnite in 2022.

Serpentine’s 2024 “Year of AI” kicked off with Refik Anadol’s “Echoes of the Earth: Living Archive,” an exhibition of works drawing on visual data of coral reefs and rainforests; it continues in October with an exhibition by Holly Herndon and Mat Dryhurst that showcases AI-abetted music-making technologies and ways of thinking about AI in a changing creative economy. The institution also published Future Art Ecosystems 4: Art x Public AI, the fourth in a series of Future Art Ecosystem(FAE) reports drawing on Serpentine-led research into technological matters at play in the art world (previous reports include Art x Decentralised Tech, Art x Metaverse, and Art x Advanced Technologies).

To survey a churning year of activity and the context from which it evolved, ARTnews spoke with Obrist about his early interest in cybernetics, his establishment of a dedicated Arts Technologies department, and his thoughts on the future of a fast-developing field.

When did you first become interested in AI-related art? How far back does that curiosity go?

I’ve been thinking about it for quite a while—in a way, back to my student days. When I started as a curator in the early ’90s, I was inspired by Billy Klüver and Experiments in Art and Technology (E.A.T.). I went to meet Billy, who with Robert Whitman started that amazing project with Bell Laboratories. I was wondering what could be a sort of E.A.T. for our time and how we could bring art and science and technology together. I was also part of the Academy for the Third Millennium and met lots of scientists through that. Actually it was Bruce Sterling, the writer who, alongside William Gibson, was one of the pioneers of cyberpunk, who got me on email in the early ’90s. At the time not many people had it, and he sent me to cybercafe to open an email account. Then I started to try to find people in the art world who were on email.

It’s important to remember that this current wave of AI is not the first. A great book by Stuart Russell, Human Compatible: Artificial Intelligence and the Problem of Control (2019), discusses different moments of AI over the decades since the ’60s. A group of people who were very pre-AI were into cybernetics, and I would say Heinz von Foerster was my entry point to it all. I became friendly with him—he was from the second generation of cyberneticians, worked with Norbert Wiener from the mid-’40s onward, and cofounded the field of cybernetics. I remember conversations with him: I was in my early 20s, he was in his 80s, and I asked him to explain to me cybernetics and the circularity. He was the first one who told me about AI and said, “this is going to be big.” At that time there was more AI optimism. Stuart Russell writes about those moments and then when it turned toward pessimism and certain moments when AI seemed not to progress. Now, of course, there is so much progress.

There was also awareness early on through sci-fi. I was always fascinated by Frank Herbert’s Dune and the unrealized film with Alejandro Jodorowsky and Salvador Dalí. I went back and read the original text and there is a commandment: “Thou shall not make a machine in the likeness of a human mind.”

Serpentine established an Arts Technologies department 10 years ago, in 2014. What was the impetus for it then, and how would you describe the early years of that initiative?

I was wondering how an institution could create something like Experiments in Art and Technology but didn’t really know how to do it. Then I gave a talk at a TED conference in Marrakech, where David Chipperfield talked about the future of museum architecture and I talked about curation. At the end of the conference, at a party, a young technologist from London, John Nash, approached me and said that he thought museums were so behind in terms of technology. He said, “How come no art institution has a chief technology officer?” That’s exactly what I had been thinking for a long time, but I didn’t know where to start. He was part of this group in London, with Dean Kissick, who is now a well-known writer in New York; Ed Fornieles, who is a well-known artist who worked early with the blockchain; and Ben Vickers, who is kind of a guru of art and technology and also worked a lot in education with universities. This group was totally fascinating, and I immediately felt with Ben that we had found the person we were always looking for, because he was both a technologist and an artist. He was a bridge, and he was interested in creating a kind of art for all with technology.

More or less 48 hours later we asked Ben to work for Serpentine. He was initially the curator of digital, [but] it took some time for the department to begin. Once the department was in place, we started to do one project after another, initially on the website. One of the very early website projects was by Cécile B. Evans: she made a bot as an early AI project that still lives on the site. We did an early online piece with Ian Cheng. Then it became clear that we wanted to bring this out into the exhibition space, because most museums would only put tech on their website. It was a bit like with video: when I began in the ’80s and early ’90s, video was still quite marginalized and didn’t have the same presence of painting or sculpture. I remember in the ’90s Bruce Nauman starting to project video big in exhibition spaces, and then came the generation of Isaac Julien, Pipilotti Rist, and other people.

Technology hadn’t yet had that kind of visibility, so we decided not only to have the Arts Technologies Department grow but also to build a bigger department of curators to allow us to produce work—not only exhibit work but produce it. That’s what’s happened over the last couple of years. Now we have the possibilities to do AI exhibitions. We are producing a project with Holly Herndon and Mat Dryhurst, opening in October as part of our “Year of AI.” Earlier this year we opened a Refik Anadol show. We’ve also produced video games. Last year, we had a show of Gabriel Massan, a young Brazilian artist who always had this dream to do a video game.

What connects all those different projects for you?

The idea that we would produce and make artists’ dreams become reality. We also want to foster new alliances with partners in technology. We did several projects in partnership with the Google Cultural Institute. We did an art and tech project with Jakob Kudsk Steensen and BTS that brought in a big alliance with K-pop. We did a project with KAWS and Acute Art and Fortnite: we were on the landing page in Fortnite and we had 150 million people who saw that show. Tens of thousands of very young people brought their parents to the gallery—normally you wouldn’t have a 10- or 11-year-old bring their parents to the gallery. It created an interest in a new generation. We also want to consider how things can be used for the public sector. This is why, besides the exhibitions, we issue our FAE reports. The idea is to support artists all over the world and also to diversify audiences.

The latest Future Art Ecosystem report is on “Art x Public AI.” How would you describe the response to that so far? Has it surprised you, or made you think differently about any of its findings?

Things always take time. It’s a slow process and over the last 10 years, we’ve worked with all these artists. With Ian Cheng we launched an AI project in the gallery [in 2018]. If you look in the visitor book and digital comments for that, all of a sudden, the tone completely changed, because some people said they came back every day because Bob, the AI character that Ian created, was so nice to them. And then some people said they were disappointed by how cold Bob was. It was like a living organism. The biggest fascination over the years with all these shows—also with Pierre Huyghe and Hito Steyerl—is that the artwork becomes a living organism with a feedback loop and a learning system. It goes back to cybernetics.

I think that that is a very major change right now in the art world. And it’s not only digitally but also analog. We have an ecology department to work with general ecology and we just had a conference about the connection that artists have to land. More and more artists are building gardens and farms. So artwork becomes a living organism but in an analog sense.

To come back to your question, it takes time to have an impact beyond the art world. The reports are useful for other museums that are now thinking about how to connect to technology. They are strategic briefings based on work by our technology team, who are interviewing artists and technologists throughout the year. All these interviews are then distilled into Future Arts Ecosystem reports. This year’s on AI, of all the reports, has so far generated the most interest. All of a sudden, everybody in the world wants to do something with AI. We can also see that with the exhibitions. The Refik Anadol show broke audience records, and there’s so much interest in our show with Herndon and Dryhurst. It took some time to build this all up, and now it’s clearly going beyond the art world.

What have you learned from artists while working on such projects?

We’ve had 10 years to learn from artists. As an institution we can never forget that the artists change us. We worked for two years with Gabriel Massan on the video game. We’ve worked for almost two years with Herndon and Dryhurst. What they are doing with music and machine-learning is super interesting because they don’t only consider what it means for their practice but what it means for the general sector. They think about new economic models, because they’re concerned about the fact that, some years ago, the music field was suddenly changed by downloads. Now AI is raising huge challenges for the economy of artists, and they are very focused on strategies for artists to assert their agency in a creative economy.

As someone who’s been interested in AI for so long, are there common misperceptions or misunderstandings that you see out in the world now that a lot more people are a lot more aware of AI and what it might mean?

Paul Klee said that art can help make the invisible visible, and I’ve been thinking about that lately in relation to AI practices that are giving visibility to something that is difficult for people to grasp. That’s also where museums can play a big role. Refik Anadol’s show worked with the Large Nature Model, which is the first open-source AI model dedicated to nature. It was hundreds of millions of images, and it allowed us to think about how to be in communion with the environment and not be separated from it. I see that as part of the function of the museum: we follow artists and show their work, but we also create an interface that makes experience possible. That’s also going to happen with Herndon and Dryhurst’s show, when the public will see what AI can do with sound and voices. It’s important that we also think about participation. It’s important that we not just explain and didactically show, but also give visitors a chance for experience. The question is not should we use AI or not use AI, but how can we use AI as a potential while ethically sourcing data without appropriating the entire life of someone not being credited or compensated in any way. We need artists in this discussion. Artists have really important things to contribute to the whole debate, so we need to put artists in the places where actual decisions are made.

Are there any other AI-related initiatives on your horizon?

Herndon and Dryhurst’s exhibition will tour in Europe and beyond Europe. That will be a lot of work, because we want the work to evolve. That’s another thing that is interesting: when we do a painting or sculpture show, an artist creates works and then maybe they go to another city. But with technology works, there usually isn’t transport involved, because it might be rebuilt physically with local [components]. And algorithms continue to evolve, so a show will be different and more advanced in the second, third, or fourth iteration.

We also want to not abandon our interest in gaming, because we think video games are really crucial. Today, more than 3 billion people globally play video games—that is more than one-third of the world’s population. The average age is 36, so that means a lot of people in their 30s and 40s and 50s play video games. It’s not only teenagers in their parents’ basements. What used to be a niche activity has become bigger than the music and film industry together. Whenever I make studio visits now, many artists—not only tech artists but others who work in more traditional mediums—have a dream to make a video game. So, we also want to continue that in 2025 very intensely.