Enter the mesmerising, AI-driven world of artist Refik Anadol

Surveying and synthesising more than 200 years of art from MoMA’s collection would be a daunting, and likely insurmountable task for most artists and researchers. Not so much for Refik Anadol, who recently unveiled a major installation in the museum’s ground-floor Gund Lobby, using AI art to generate endlessly changing forms and sounds across a 24ft x 24ft media wall, based on 320,000 visual inputs.

Unsupervised, as the installation is tagged, is a major career moment. ‘To show at MoMA is one of my biggest motivations in life,’ describes the Turkish-born, LA-based media artist. But numerically speaking, it is far from the most ambitious. In 2019, he’d used 100 million photographs of New York City, found publicly on social networks, to create a 30-minute cinematic piece. For a 2020 exhibition at Melbourne’s National Gallery of Victoria, Anadol deployed Google AI’s algorithms to process around 200 million nature and landscape images to create a 3D visual piece, Quantum Memories. The following year, his contribution to the Venice Architecture Biennale, Sense of Space, involved a collaboration with neuroscientist Taylor Kuhn to develop machine-learning algorithms based on 70 terabytes of MRI data, then used it to imagine the development of brain circuitry throughout the human lifespan. Not only is Anadol fascinated by what data sets tell us about the world, he also uses words like ‘beautiful’ and ‘inspiring’ to describe them.

‘In 2008, I coined the term “data painting” to express the idea that data can become a pigment that reflects imagination. This has driven my practice for 14 years,’ he tells me via Google Meet from his LA studio. Rather than conventional paintings, which involve fixing paint on canvas, he creates ‘living paintings’ which morph and evolve infinitely.  

‘I became the first artist-in-residence at Google in 2016, which made me realise that a machine can learn, it can remember, and it can dream,’ Anadol continues. ‘Machines are becoming part of our society, and now they’re in our creative practice as well. It’s a whole new world.’

He explains that the MoMA installation has its origins in a 2021 online exhibition on the digital art platform Feral File, for which he’d trained a machine-learning model to interpret publicly available visuals and information around the museum’s collection to create a piece of generative art. The result was MoMA’s first NFT collaboration, ‘reimagining the trajectory of modern art, paying homage to its history and dreaming about its future’. ‘The idea is to stand on the shoulders of these incredibly pioneering artists [in the museum’s collection] to create something new,’ says Anadol.

Unsupervised takes the collaboration further, introducing a physical dimension as well as incorporating live inputs. Sensors within the Gund Lobby – which detect changes in light, numbers and movement of people, as well as the weather – will inform the visuals and sound, further reinforcing the idea of a living artwork. It is true that technology will be doing a lot of the heavy lifting, but Anadol is keen to emphasise the importance of human involvement. There is a lot of work that his 15-strong team (including computer graphic experts, architects, designers, musicians, data scientists, and AI researchers, who all together come from 11 countries) have put into creating a new AI model, and setting parameters so the AI can make decisions around colours, forms, patterns and speed. ‘It's not an autonomous piece, because I don't believe that’s what the future should be. I think human and machine collaborations are more relevant and positive for the future.’

The installation was co-curated by Michelle Kuo, curator of painting and sculpture at MoMA, and Paola Antonelli, the museum’s senior curator of architecture and design, and director of research and development. ‘With this commission, MoMA underscores its support of artists experimenting with new technologies as tools to expand their vocabulary, their impact, and their ability to help society understand and manage change,’ explains Antonelli.

Certainly, seeing Anadol’s digital animations come to life on the media screen in the Gund Lobby drives home the importance of physical experiences in our increasingly digital age. It puts Unsupervised in a long line of projects where Anadol has brought digital art to architecturally significant spaces, such as the façades of Frank Gehry’s Walt Disney Concert Hall in LA, Zaha Hadid’s Dongdaemun Design Plaza in Seoul, and Antoni Gaudí’s Casa Batlló in Barcelona. ‘I enjoy the moment when the physical and virtual connect. It’s always powerful to bring the two dimensions together,’ says the artist. His one-night-only projection mapping performance at Casa Batlló, in May 2022, had drawn 48,000 attendees. He spent the night walking among the audience, many of whom came up to him, moved to tears and asking for hugs. ‘That moment, when you touch someone’s mind and soul, and trigger beautiful emotions, is the ultimate moment of success,’ he reflects.

Up next for Anadol is a new take on the metaverse, called Dataland. As the project’s placeholder describes, it’s ‘the world’s first multi-sensory metaverse project [...] we will architect unprecedented spaces and invent cutting-edge poetic algorithms for new meditative experiences in the metaverse’. Collaborators include leading neuroscientists, architects, and AI pioneers, as well as tech titans such as Nvidia, Google and Epic Games. ‘I want to show that the metaverse is not just a virtual, cold space with heartless and soulless machines,’ he explains. ‘This is our attempt to find narratives in this new galaxy of imagination.’

‘Refik Anadol: Unsupervised’ is on view until 5 March 2023 in the Gund Lobby at MoMA, moma.org (opens in new tab); refikanadol.com (opens in new tab); dataland.art (opens in new tab)

Sample data visualisations of Refik Anadol’s Unsupervised — Machine Hallucinations —MoMA — Fluid Dreams (2022) appear on the limited-edition subscriber cover of  January 2023 Wallpaper*, ‘The Future Issue’. Featuring real-time digital animation on LED screen and sound, the data sculpture was created using custom software and a generative algorithm with artificial intelligence.

A version of this story appears in the January 2023 issue of Wallpaper*, available now in print, on the Wallpaper* app on Apple iOS, and to subscribers of Apple News +. Subscribe to Wallpaper* today (opens in new tab)