The Jeu de Paume art centre has plunged into artificial intelligence with Le Monde selon l’IA (“The World According to AI”), on view from April 11 to September 21, 2025. This ambitious exhibition combines art, photography, cinema, literature, and music to examine how creative practice intersects with new technologies critically and experimentally. Rather than a tech demo of flashy algorithms, the show curated by Antonio Somaini with associate curators Ada Ackerman, Alexandre Gefen and Pia Viewing interrogates the cultural phenomenon of generative AI at a moment when it inspires both excitement and anxiety.
Spread across the museum’s entire gallery space, Le Monde selon l’IA is organized around two overarching strands of artificial intelligence. It moves from “analytical AI” systems, such as facial recognition and computer vision, to “generative AI” models that produce images, text, or sound. This structure reflects the exhibition’s intent to highlight the evolution of contemporary AI, from tools that analyze the world to those that create new content. The curators use this division to frame key questions: how do machines see the world, and how might they reshape artistic creation?
Visitors navigate a trajectory contrasting AI’s surveillant, data-crunching side with its creative, image-making capacities. Throughout, the curatorial approach remains notably didactic and panoramic. Interspersed “time capsule” stations inspired by cabinets of curiosities punctuate the route with historical context. These mini-displays present early computing devices, vintage attempts at machine vision, and other milestones that trace a genealogy of AI, underlining that today’s machine-learning feats emerge from decades (even centuries) of technical and cultural developments. This blend of contemporary art and archival material embodies the curators’ strategy of linking present innovations to past imaginaries. It also gives visitors reference points to grasp better the radical newness and long roots of 21st-century AI.
By structuring the show around analytical and generative AI, the curators address a spectrum of social, political, and aesthetic concerns. On the “analytical” side, surveillance and algorithmic bias issues come to the fore. For example, Trevor Paglen’s “Faces of ImageNet” installation exposes the troubling biases baked into facial recognition training data, spotlighting how AI classification can reinforce prejudices. Nearby, Kate Crawford and Vladan Joler’s “Calculating Empires” presents a sprawling research-based artwork charting five centuries of inventions leading to modern AI. This visual genealogy reveals how today’s algorithms are built on historical systems of power and labour, implicitly asking who benefits and who is exploited in the age of AI.
On the “generative” side, the exhibition embraces AI’s creative and even uncanny outputs. Several projects demonstrate how artificial neural networks autonomously produce texts, images or music, raising questions about authorship and originality. One such work by musician Holly Herndon and collaborator Mat Dryhurst invites viewers to engage directly: their video piece “xhairymutantx” is composed of bizarre, fantastical imagery synthesized by a custom AI model, and it lets visitors scan a QR code to prompt the AI themselves. This interactive element exemplifies the exhibition’s experimental ethos, blurring lines between artist, audience and algorithm.
Across the gallery, other contributions by artists like Hito Steyerl, Agnieszka Kurant and Christian Marclay further probe AI’s impact on how images and knowledge are produced. From dataset-driven plant “portraits” to AI-remixed soundscapes, the works on display illustrate both the promises and perils of delegating creative decisions to machines. Together, they speak to AI’s dual identity as a surveillance and control tool and a new medium of expression.
Le Monde selon l’IA arrives at the intense public debate about intelligent algorithms, and the exhibition positions itself squarely within this discourse. In a cultural climate grappling with AI ethics, data privacy, labour automation, and image-making’s future, the show’s content resonates more than art-world contemplation; it’s a forum for examining urgent questions. The curatorial team reinforces this relevance by extending the exhibition beyond static artworks. A program of film screenings and talks with experts runs in parallel. At one point, the museum even stages a mock “trial” of an AI, putting the technology in the dock for society’s scrutiny. Such programming underscores the exhibition’s journalistic tone, treating AI not as science fiction but as a civic matter to interrogate. A comprehensive bilingual catalogue, with essays from researchers, further situates the artworks within broader ethical and philosophical contexts. Crucially, the setting of Jeu de Paume (an institution devoted to photography and image-based art) lends an added layer of meaning. Many pieces address how AI systems generate or interpret images, directly challenging traditional notions of photographic truth and artistic agency. By hosting this exhibition, a venue known for chronicling the evolution of visual media acknowledges that the new frontier of image-making may no longer involve a camera at all but code and computation.
