France: three senators ready to reverse the burden of proof on AI training data
In the wake of the failed dialogue between rights holders and AI developers, three senators are preparing an explosive proposal: introducing a presumption that copyrighted works were used to train AI models. A move that could reshape the balance of power with tech platforms and push the limits of the European legal framework.
A parliamentary initiative could profoundly alter the balance of power between creators and developers of generative AI on the French territory. According to our information, senators Agnès Even, Laure Darcos and Pierre Ouzoulias – all three rapporteurs of the information mission « Creation and AI » whose conclusions were delivered in July – are preparing to table, in the first quarter of 2026, a bill aimed at reversing the burden of proof in terms of the use of protected works for training generative AI models.
Bensamoun Report
The idea is simple, but its implications are far-reaching. Today, when a right-holder goes to court on suspicion that a model has been dragged onto his or her work, it is up…
Enjoy unlimited access to our articles:
Subscribe nowAlready subscribed? Log in here. Connectez-vous ici.
Follow us on Twitter: @3l3ctr0nLibr3