Translated by Bethszabee Garner
At first glance, it is a mountain landscape. Mont Blanc stands out in an almost unreal light, the peaks bathed in a digital fog that seems to breathe. But very quickly, something goes wrong. Human silhouettes stretch and distort, and visual artifacts appear: geometric rectangles, floating pixels. Under the hand of Jacques Perconte and the eye of upscaling algorithms, nature becomes digital matter in motion. These algorithms, designed to “invent” additional pixels to improve image definition, interpret and sometimes distort the initial image.

The result is a semi-digital landscape, halfway between real capture and algorithmic reinterpretation. The image is at once sumptuous and unstable, hypnotic and artificial. Here, artificial intelligence does not seek to imitate reality: it deconstructs it, interprets it, and ultimately exposes it as a technical process in its own right. A poetry of flaws and imperfections that sums up the central issue posed by AI in contemporary artistic creation: when machines generate works of dizzying mastery, what becomes of the artist's place? A mere machine operator or co-author of a hybrid imagination?
New medium
These questions are stirring up the art world and society as a whole. For Antonio Somaini, the answer is clear: “Absolutely not.” But he immediately qualifies this by saying, “It all depends on the sector concerned.” In fields such as illustration and advertising, AI models are already being used to produce visuals at lower cost, leading to a worrying level of automation. On the other hand, in the field of contemporary art, galleries, and museums, AI appears more as a new medium than a substitute for the artist. “Artists remain present at every stage: choosing the model, setting the parameters, selecting the results...,” he insists.
The Jeu de Paume exhibition illustrates this diversity of approaches. Some artists, such as Nora Al Badri, adopt a critical and political stance. Others, such as Grégory Chatonsky, question the very notion of creation, pushing AI to generate cities, faces, stories, and memories in post-human universes. “The exhibition brings together both young artists and more established figures: Nora Al Badri is in her forties, Chatonsky already has a long career,” points out Antonio Somaini.
“We are entering a form of cohabitation, of co-creation, which is becoming increasingly common. I also see this at the university where I teach: students now co-write with AI, which raises new pedagogical and ethical questions,” he continues.
Algorithmic sensitivity?
Far from being confined to cold mechanical reproduction, AI is also becoming a generator of sensitive material. “These models generate sensitive material: images, sounds, texts. We often perceive a kind of strangeness in these productions, an algorithmic otherness. But this strangeness can itself be a source of emotion,” explains Antonio Somaini.
It is precisely this otherness that Egor Kraft works with when he reconstructs imaginary ancient artifacts, or Nora Al Badri when she liberates the data from the bust of Nefertiti. “If the use is transparent, it allows us to explore possible pasts, fictional variations of reality, with real aesthetic power,” continues the curator.
But behind this algorithmic emotion, the logic of standardization of large AI platforms also comes to the fore. “The challenge is to escape the standardization imposed by large AI platforms, which often produce stereotypical and repetitive images. Artists, on the contrary, explore the field of possibilities beyond these clichés,” insists Somaini.

But the exhibition also shows the material downside of these technologies: works such as those by Hito Steyerl and Agnieszka Kurant denounce the invisible workers behind the click and the resources devoured by data centers. For beneath the apparent magic of the images generated, a whole technical, economic, and political ecosystem is taking shape.
It is this hidden side of digital technology that emerging Franco-Swiss artist Julian Charrière places at the heart of his work. Interviewed on France Culture's Le Cours de l'histoire program (March 2024), he points out that "we often talk about artificial intelligence as if it were floating in the cloud, but this cloud is made of concrete, water, and mining. The beauty of AI-generated images is based on very concrete, very physical supply chains." Through his approach, the artist reminds us that behind the apparent fluidity of data flows, there are very real infrastructures, disrupted territories, and exploited resources, right down to the bowels of the Earth.

Human creation is augmented, but not dissolved... Through these multiple works, the exhibition “The World According to AI” rejects any simplistic view. The artist is not supplanted by the machine: he redefines his gesture in contact with it. "We must get rid of the idea that there is a uniform AI. On the contrary, there is a plurality of models, logics, and modes of operation. The exhibition shows this diversity," concludes Antonio Somaini. Faced with this new grammar of the senses, co-creation becomes a laboratory of hybridization. Artists manipulate AI, playing with its errors, its strangeness, its biases. A new creative scene where, far from being a mere spectator, the artist remains the author of the disturbance.

Exhibition “The World According to AI,” Jeu de Paume (Paris), until September 21, 2025.
An article written by Désirée De Lamarzelle, featured in issue 12 of OniriQ magazine.



