French Publishers Sue Meta Over Alleged Copyright Violations in AI Training

Meta is accused of using copyrighted texts from 200,000 books, including French works, to train its AI models without proper authorization.

Three French organizations representing publishers and authors have filed a lawsuit against Meta for alleged copyright violations in the development of its AI model, Llama.
The lawsuit claims Meta used a database, Books3, containing texts from approximately 200,000 books, including French works, without obtaining permission from authors or publishers.
Meta previously acknowledged using the Books3 database in a U.S. lawsuit in 2024, defending its actions under the 'fair use' doctrine, which is not universally recognized.
The plaintiffs argue that Meta's practices harm the cultural sector and demand accountability and compensation for the unauthorized use of protected works.
French publishers and authors stress the need for AI companies to respect intellectual property laws and establish fair compensation mechanisms for creators.