French National Assembly Blocks AI Copyright Bill

A French bill that would have shifted the burden of proof onto AI providers to demonstrate that copyrighted content was not used in training their models has been blocked by the National Assembly. The Senate had adopted the bill on April 8, 2026, after amendments incorporating observations from the Council of State, which issued a favorable opinion on March 19. The bill, introduced by Senator Laure Darcos on December 12, 2025, aimed to protect copyrighted works used as input for generative AI systems. It proposed a rebuttable presumption of use, triggered by evidence such as AI outputs or technical reports, and applied only to civil litigation. Despite these safeguards, the bill was not placed on the parliamentary agenda, effectively stalling it. The outcome reflects intense lobbying by tech companies including Google, Meta, and Mistral, who argued for innovation and competitiveness. Cultural stakeholders, represented by collective management organizations like SACEM, SACD, SCAM, and ADAGP, now lose a key legislative tool but may pursue legal avenues under existing French and EU law.