Not only GTC, Refik Anadol, a Turkish-American new media artist, has captivated audiences worldwide with his work at the intersection of art and artificial intelligence. He studied photography, video, and fine arts in Istanbul before moving to the United States to pursue a master’s degree in design media arts at UCLA. 

On the sidelines of NVIDIA GTC 2024, Anadol discussed his latest project, ‘AI Nature Model’, which transforms roughly 750,000 images of flora, fauna, and fungi to create visuals evocative of an Amazonian rainforest. “I believe we may need a new perspective because current AI research focuses on human intelligence and human reasoning, which is very important, but I think nature needs a new perspective,” Anadol explained.

He transformed the Las Vegas Sphere, a 580,000-square-foot dome blanketed with programmable LED panelling, into the world’s largest AI artwork. His feat of art and engineering, called Machine Hallucinations, was specifically designed to take advantage of the structure’s unique architecture. 

Anadol’s work has been exhibited at prestigious institutions such as MoMA in New York, where his installation ‘Unsupervised’ became the first generative AI artwork to be collected by the museum. Anadol revealed, “The artwork was running on an extremely complex system, and the NVIDIA DGX station was behind the scenes running the artwork all the time, ever-changing.” 

The artwork’s popularity led to extended display times and attracted nearly 3 million visitors, with his shows being fully booked. 

Collaboration has been key to Anadol’s success, with partnerships spanning across hardware, software, and data providers. He said, “Last eight years, thanks to our very first collaboration, I guess, with Gamescom and many others, we have a lot of local and physical support. DGX stations, and A6000 recently – all these gears are helping us so much in working locally with big data.” 

Further, he also credited NVIDIA’s software ecosystem, including NVIDIA Omniverse and Picasso, for enabling the collection, cleaning, sorting, and curation of half a billion images of open-source nature data from institutions like the Smithsonian, Natural History Museum, and National Geographic.

Anadol’s GTC 2024 installation took visitors on a multi-sensory journey, with AI dreaming of the flowers of the Amazon and allowing visitors to smell the possibilities of new AI dreams. Anadol explained, “I believe that scent has a profound experience on our life as a memory form,” highlighting his vision for the future of AI art that goes beyond just text, image, and video.

As for those who may be apprehensive about generative AI, Anadol emphasises the importance of understanding the medium. He said, “I think the best answer is really understanding the medium because as an artist myself, like the last eight years collecting, curating, and sorting data, training models, I think this is one of the best ways of owning the narrative of the work, and it creates more possibilities.”

AI Art Beyond Prompt Engineering

While Refik Anadol’s work showcases the stunning possibilities of AI in art, other artists are exploring the technology in diverse and innovative ways. 

Sougwen Chung, a Chinese-Canadian artist, blends art and AI through robotic collaborations. In her project ‘Drawing Operations Unit: Generation (DOUG) 2’, Chung trains AI models on her abstract paintings and uses them to control robots that paint alongside her on large canvases. As Chung explained, “What I’m chasing is that surprise and wonder in machine translation,” creating a feedback loop between artist and machine.

Trevor Paglen, another pioneering AI artist, employs artificial intelligence to critically examine surveillance, data collection practices, and the invisible infrastructures of the digital age. He said, “What’s interesting to me is looking at AI as not only a set of technical systems, but systems that have culture built into them, that have politics implicitly built into them and trying to unpack that.” 

Paglen’s work ‘ImageNet Roulette’ collaborates with AI researchers to uncover biases in facial recognition datasets. 

These artists demonstrate that AI art is not merely about crafting the perfect prompt. It is a multifaceted relationship between humans and machines, data and creativity, and the ethical implications of AI in the art world. 

The integration of AI in art is not limited to the traditional gallery space. As Refik Anadol mentioned, the future of AI art may include interactive experiences that engage multiple senses, such as touch, taste, and smell. This ‘generative reality,’ as Anadol calls it, hints at a world where AI-generated characters and environments become increasingly immersive and tangible.

Anadol encourages experimentation with existing tools and platforms for aspiring AI artists, from RunwayML and Stable Diffusion to Midjourney and OpenAI’s GPT models. By understanding these systems and their capabilities, artists can better navigate the complexities of AI art and develop their own unique approaches to the medium.