Jeroen van Loon NL
Reinforced Learning

Video installation

Exhibition : “Disobedience in the Digital Age”

From October 23 to December 13, 2026
From Wednesday to Sunday, 2pm to 7pm

Friche La Belle de Mai
, La Tour, 4th floor
, Marseille

Credits

Cross-TIC Co-production

“Reinforced Learning ” is a video installation that combines humor and critical reflection to challenge future scenarios related to robotization and AI. It uses AI specialized in visual effects to replace humans with robotic humanoids in various “blooper videos.” The result is a series of videos in which it is not us, humans, but this time the robots, who engage in irrational and ineffective actions. They depict a strange future in which robots mimic our illogical behavior and challenge our widespread belief in their non-human capacity to achieve perfection. The title *Reinforced Learning* comes from the field of machine learning, in which algorithms learn by imitating specific actions. The algorithm receives a reward when it successfully mimics the intended action. This work depicts robots learning from “bad” input: human imperfection.