In the film "Chronicle of the Black Sun" by artist Gwénola Wagon, the AI ​​chooses the image of a young woman to mask a sun that has become scorching.

Until September 21, the Jeu de Paume presents “ The World According to AI », the first collective exhibition of this scale, bringing together forty artists from the French and international scene - including Julian Charrière, Christian Marclay and Trevor Paglen - with around fifty works produced since the mid-2010s. The visitor is drawn into it as if on a treasure hunt: from one work to another, through artistic experiences and the questions they raise, theartificial intelligence reveals itself, for better or for worse. Far from any technophile fascination, as well as its technophobic undertone, the exhibition offers a detailed analysis of the way in which AI shapes our relationship to images, to others, to society. It is also a free critique, carried by artists who, far from being dominated by the power of the tool, seize it to play with it, thwart its flaws, and help us, as users, to better sketch the contours of this new world. Guided by the exhibition curator, academic Antonio Somaini, we attended the inaugural tour of the itinerary. Here are the four highlights.

Excerpt from the work “What do you see, YOLO9000?”

1/ Preamble

' " The World According to AI » brings together works by artists who, over the past ten years, have responded in multiple ways to the growing presence of artificial intelligence in all strata of society, culture, and the economy, mobilizing different strategies and diverse mediums. The exhibition brings together nearly forty artists—including ten French artists—for around fifty works, ten of which are presented for the first time. The World According to AI » thus constitutes the first major collective exhibition devoted to artificial intelligence since 2022, a pivotal year marked by the launch of ChatGPT and the models text-to-image like MidJourney, which inaugurate a new typology of AI called " generative ", capable of producing texts, still or moving images."

2/ Mapping AI

Kate Crawford & Vladan Joler, Calculating Empires 2023

“Science is matter! In this first part of the exhibition, we wanted to highlight this material—even geological—dimension of digital technologies, in particular that of artificial intelligence through its environmental impact. Because these are in no way dematerialized technologies, contrary to what the metaphor of “ cloud " These systems consume immense amounts of energy and natural resources.

This work, entitled " Anatomy of an AI System ", is a spatial diagram created in 2018 by Australian researcher Kate Crawford and Serbian researcher and artist Vladan Joler. It is a map of artificial intelligence, highlighting the vast network of extraction and exploitation of material resources, human labor, and data, necessary for the functioning of this small black box displayed next to it: Alexa. Launched by Amazon in 2016, Alexa is a voice-controlled artificial intelligence system, present in many American homes. She can give the weather, play music, turn on the lights... " Anatomy of an AI System » thus contrasts the apparent simplicity of the object with the complexity of a heterogeneous entity, where theories, technologies and ideologies intersect.
A second diagram by Kate Crawford and Vladan Jolen is featured in the exhibition: “ Calculating Empires » (2023). This is a masterful work that required four years of research. This time, the aim is to retrace the techniques and systems of thought that gave rise to artificial intelligence, since the Renaissance and over five centuries, and to understand how it developed to its current forms. The work extends over entire sections of walls. According to the artist duo, generative AI thus draws its source from the linear perspective of the Renaissance - the first mathematization of space - all digital images being, they point out, composed of grids of pixels defined by coordinates. The diagram also goes back to the origins of ChatGPT: the invention of printing and the first forms of text reproducibility.

3/ Generative AI and latent spaces

Egor Kraft Content Aware Studies 2019

Nouf Aljowaysir Man In Arab Costume 2020

"After a first part devoted to analytical AI, the second section of the exhibition explores generative AI and the notion of " latent spaces ", encoded data matrices, texts and images. Since the mid-2010s, several contemporary artists, like those presented here, have taken over these spaces to project dystopian futures, revisit the past or construct fictional memories. This is the case of the Russian artist Egor Craft (born in 1986, based in Saint Petersburg), who has taken an interest in fragments of ancient Greek sculptures. In " Content Aware Studies ", he trains an AI model from thousands of 3D scans of sculptures and friezes to generate elements to replace the missing parts. AI doesn't just restore missing objects; it also creates new forms, like these distorted faces, the result of an attempt to imagine a past that never existed. This work echoes an 18th- and 19th-century tradition of engravings of prints that aimed to reconstruct the ancient city of Rome.

Justine Emard, for its part, mobilizes AI in its video projection " Hyperphantasia »: the French artist trained a model on data from the Chauvet Cave to generate new variants of prehistoric drawings and sketch out an alternative prehistory. These works invite us to renew our perspective on the mystery of the origins of art. Finally, the Saudi-born artist Nouf AlJowayzir (born in 1993) is interested in the limits of analytical AI when it comes to representing non-Western populations. Indeed, the models are mainly fed by Western images, and reveal significant biases in their ability to recognize or reproduce historical figures, such as those of the Bedouins or other Middle Eastern peoples.

4/ Generative writings

"Presented for the first time in French, the work " Rewrite » by Canadian poet David Jhave Johnston, like the other pieces exhibited in this room and the next, questions the modalities of literary collaboration between humans and artificial intelligence. What are the limits and impacts of AI on writing practices? Jhave Johnston has an AI generate a text, which he then modifies in real time. He describes this process as " AI flow sculpture ", a way of highlighting the intrusion of artificial intelligence into our lives, and the need to shape this new matter. AI appears here as a partner: a collaborator in our ways of thinking, living, and writing. An accomplice of poetic expression.

Linda Dounia Rebeiz extends this reflection on language and the mystery of writing produced by AI. It understands all languages, can switch from one to another, and even seems capable of interpreting languages ​​it has never encountered. As a universal translation tool, it comes to repair the trauma of the Tower of Babel. But ultimately, what language does it speak to itself? In her work " flip flops », which combines traditional calligraphy and artificial intelligence, Linda Dounia Rebeiz poses this question of mystery: that of the language specific to the machine.

FR Jeu De Paume Gregory Chatonsky The Fourth Memory 2025

Dive alongside the artist into a post-apocalyptic world where only AI would have survived.

“This installation, made up of images, videos, and sculptures generated by AI, is a self-portrait of the artist I dreamed of being, but did not become during the 20th century. To do this, I fed several AIs with personal documents, my photos, my voice, my texts... The screen continuously projects a video that shows everything I have not lived: all my possible lives. My life as a man, a woman, or with a different skin color... The sounds are also generated artificially: it is my cloned voice that we hear, a voice close to mine, but not quite the same. The visible character also resembles me a little; the machine imagines multiple versions of myself that seem to explore uncertain places. I wanted this work to raise the question of the extinction of life. Artificial intelligence appears at a time when the survival of our species is becoming uncertain. Through her, I ask: what is a tomb when there are no more witnesses?