Egor Kraft, Content Aware Studies, 2019 © Egor Kraft
How do contemporary artists perceive the rise of artificial intelligence in our lives? With “The World According to AI,” the Jeu de Paume offers a critical and poetic reading of this revolution, through the works of some forty artists. We met with Antonio Somaini, the exhibition's curator, who sheds light on the human, cultural, and aesthetic issues raised by these new forms of creation—ranging from co-creation, collective memory, and generative hallucinations.
Can you tell us about the exhibition and its main intentions?
The exhibition “The World According to AI” at the Jeu de Paume presents the work of some forty contemporary artists who, over the past ten years, have responded to the growing presence of AI across all levels of our cultures, our society, the economy, and work.
The intention of the exhibition is to help us understand what it means today to experience the world according to AI or through AI.
How are artists reacting to this new reality?
The exhibition shows how artists have responded to this new condition, either to try to reveal the functioning of these systems with their biases – gender, race – their political and ethical issues, or to explore these new generative models which allow new forms of collaboration between the human and the non-human.
Does the exhibition challenge the idea of autonomous AI?
The exhibition highlights in several ways the fact that AI is not a completely autonomous or dematerialized system. On the contrary, what we call AI is entirely pervaded by human presence, natural resources, and energy.
It really shows that AI is a collective entity with multiple human and non-human contributions.
How are generative AI models transforming artistic creation?
With the spread of generative AI models, the forms of artistic creation are also changing. Artists are learning to collaborate and co-create with AI models, which are endowed with different forms of autonomy.
They explore these models through prompts, through images. They enter into forms of co-creation and sometimes even co-hallucination, in the sense that they explore the errors of these models – the “glitches,” or hallucinations.
Could you give us an example of a work exhibited?
Some artists like Grégory Chatonsky, whose installation we see here, question these models as vast repositories of possible images... Generative AI here becomes a source of virtualities, of possibilities that can be explored through its models.
What are the risks posed by exposure?
The exhibition highlights the dangers that AI poses to humanity: energy consumption, labor exploitation, job losses.
But it also explores the potential of AI in terms of co-creation, exploration of models and opening up to new forms of imagination.
Do some artists offer another vision of AI?
Yes. For example, Holly Herndon works for a different type of AI: collective, collaborative, transparent, and based on remuneration for the works of the artists who participate in its training.
Artists can really help us navigate these profound transformations we are experiencing.
What are the major human issues addressed?
One of the major questions being asked is what will become of humans in their relationship with AI. The artists explore how humans are being redefined by these technologies.
Some works link AI to other forms of collective intelligence, such as in nature – for example, a termite mound cast by Agnieszka Courrante.
And what about AI errors? Are they addressed as an artistic subject?
Yes, these flaws—or “hallucinations”—are moments when models fail to meet expectations. Many artists explore them for their poetic potential.
They often use older, less “aligned” versions of models to preserve these creative slippages.
How is the exhibition organized?
It is structured into several sections: the materiality of technologies, the mapping of AI in time and space, its relationship to other forms of collective intelligence, artificial vision, micro-work, etc.
Another part is dedicated to generative AI, which produces new texts, images and voices.
Does the exhibition offer a historical perspective?
Yes, she emphasizes that these transformations have ancient roots: linear perspective during the Renaissance, the invention of printing... Through “time capsules,” she establishes links between the present and the past.
Can we say that AI is built on human and non-human collective intelligence?
There are, of course, the programmers, the click workers, the data produced by all of us… It's a profoundly collective system.
What role does memory play in works with AI?
Some artists, like Chatonsky with The Fourth Memory , approach AI as a process of reprocessing our cultural memory, even in the event that humanity has disappeared.
What was your role in the exhibition?
I was invited as a curator. I come from an academic background and have been working for several years on the impact of AI on images, visual culture, and contemporary art.
Finally, how would you define AI?
I think AI is a machine that encodes the world – texts, images, phenomena – and then understands or imagines it.
Bringing together some forty artists, The World According to AI offers a rich and engaging journey exploring the impacts of artificial intelligence on our imaginations. On view at the Jeu de Paume (Paris, France) from April 11 to September 21, 2025.
Nicolas Sarazin