Through a deep learning process, a GAN (Generative Adversarial Network) algorithm determines the most frequently repeated motifs in collections and becomes able to predict and create new works in the plastic continuity of existing ones. On the one hand, the notions of “artist”, “conservation” and “curator” are questioned: future additions of works to a collection can now be decided by artificial intelligence. On the other hand, we have discovered that, in this generative process, the most interesting are the errors of the system itself: failed tests, intermediate tests, the technological unconscious that arises through accidents. We are then witnessing collisions between the force of chance and the art of prediction. Finally, it desecrates the old paradigm of creativity and art: we can only imitate previous gestures, never original. The meaning of an image does not lie in its origin but in its destination. So, to a certain extent, everything is “déjà vu”. (made in collaboration with Joan Fontcuberta)

This is a unique website which will require a more modern browser to work!

Please upgrade today!