Gregory Chatonsky

Born in 1971 in Paris. Lives and works in Paris and Montreal.


Recursive neural networks, usually referred to as AI, allow media to mutate. From huge databases accumulated on the Web, it is possible to produce resembling images: we provide a neural network with thousands of images of birds and new images appear that we recognize as birds even if these do not exist.

These are media media that constitute a real break in the history of the production of realism. MUE invites us to the permanent metamorphosis of images taken in the permanent flux of mutation. An alternate world becomes visible that looks like the one we know but differs from it, like the possibility of an unexplored branch. These are realistic images, but they are no longer real. They are no longer based on the capture of light (photorealism) but on the synthesis of memories captured in the nets of massive data network.

What happens to the image when it is taken in an infinite series? What is its form when it is only the form of its change? What era of realism do we enter when we recognize as realistic images that do not come from the capture of reality but from the hypermnesia of the network?