AI is on the rise in the art world. MoMA is the newest digital art exhibit to display its capabilities. The AI training comes from 180,000 works of art ranging from Warhol to Pac-Man. The installation occupies the museum’s ground-floor Gund Lobby; manifesting the integration of the nature of art and the formation of a machine’s “dream state”.


AI is quickly gaining popularity among artists. The collection, by artists Refik Anadol, “Refik Anadol: Unsupervised”, uses artificial intelligence to create new images in real-time. In detail, it processes 380,000 images of 180,000 art pieces from MoMA’s collection to create a stream of moving images. The content comes to life by a continuous change between fantastic images materialising and morphing into each other.

Credit: Refik Anadol/Courtesy MoMA

Installing a 24-by-24 foot digital display, “Unsupervised” renders an infinite animated flow of images. Each of them is dreamed up as you watch by an AI model fed by the museum’s entire collection of artwork. The flow of the collection bases itself on what happens around it, making the piece feel like it’s real and alive.

Meanwhile, Anadol states that

“I’m trying to find ways to connect memories with the future… and to make the invisible visible”

and MoMa also says in a press release,

“A singular and unprecedented meditation on technology, creativity, and modern art … reimagining the trajectory of modern art, paying homepage to its story, and dreaming about its future.”

A Breakthrough for AI Art

Credit: Courtesy Refik Anadol

The Turkish artists describe the work as a “Machine Hallucination” that brings a

“self-regenerating elements of surprise to the audience and offers a new form of sensorial autonomy via cybernetic serendipity”

There is a method with which current AI’s learn: Supervised AIs – are trained using data tagged with keywords. That means those keywords allow AI to organise the clusters of similar images, and will process new images based on what it learned.

However, with the work of Anadol, the AI was left to make sense of the entire MoMA art collection on its own, without labels. That is also the reason to explain why the title of the exhibition is “Unsupervised”.

Another important achievement of this collection, is that “Unsupervised” can generate each image in real-time. This is based on the motion of visitors and weather data of Manhattan.

“AI-generated art has arrived”

Says Brian Caulfield, blogging for NVIDIA, whose StyleGAN forms the basis for Anadol’s AI.

Much More is Coming

Anadol and the exhibit creator make us think that the art world is in a new ‘renaissance’, and that Unsupervised is the perfect example to prove it. It’s not just an exploration of the world’s foremost collection of modern art,

“but a look inside the mind of AI, allowing us to see the results of the algorithm processing data from the collection, as well as ambient sound, temperature and light, and dreaming”

the artist notes.

Meanwhile, modern generative AI models are showing unprecedented evolvement in generalising beyond particular subjects. This includes images of human faces, cats or cars. And then, AI can encompass language models that allow users to specify the image they want in natural languages, such as inpainting.

“Ultimately, generative AI has the potential to unlock the creativity of everybody from professional artist, like Refik, to hobbyists and casual artists, to school kids”  David Luebke said – vice president of graphics research at NVIDA.

Where to Watch Unsupervised

Get lost in Anadol’s hypnotic forms as the exhibition displays at MoMA NYC until March 5, 2023.

Subscribe to FIB’s Weekly Breaking News Report for your weekly dose of music, fashion and pop culture news!