French Publishers Sue Meta Over Unauthorized AI Training Practices

French writers and publishers have sued Meta, alleging the tech giant has used copyrighted work without consent to train its artificial intelligence (AI) systems. The lawsuit, filed by leading publishing and authors’ associations in France, is one of the biggest tests so far of Meta’s practices in the AI arena.
The lawsuit, filed in a Paris court, accuses Meta of using copyrighted content extensively to enhance its AI models, including those behind popular platforms like Facebook, Instagram, and WhatsApp. The plaintiffs argue that Meta’s actions amount to a form of “economic parasitism” where the company benefits from others’ intellectual property without permission or payment.
This is the first such case in France but follows others that have surfaced in the United States. U.S. authors and other copyright holders have sued Meta and other tech firms for their use of copyrighted material to train artificial intelligence. The French action is seen to be a trend-setting case with potential broader implications about how AI companies handle copyrighted material.
The charges laid against Meta are illegal use of books and other written content for training its big language models such as Llama. The plaintiffs say the actions of Meta are akin to theft, with one official labeling the scenario a “David versus Goliath” situation. The litigation aims to establish a clear precedent regarding the use of copyrighted content for training AI.
In developments, Meta is also legally in the crosshairs in the U.S. for the same allegations. A federal court allowed a case to proceed, alleging Meta stripped copyright management information from works it used to train its AI models. This is alleged to have been done to conceal the use of copyrighted works and render the AI incapable of producing information that would make the source of its training material transparent.
The U.S. case involves authors who sue Meta for alleging the company made use of their books without authorization, credit, or compensation. According to the suit, Meta copied some of the books from illegal websites, including torrent sites, and removed copyright marks to avoid detection. The case could have Meta face significant financial fines and can set a new legal precedence for AI businesses to make use of copyrighted work.
Today’s lawsuits against Meta highlight the escalating tension between tech companies and owners of copyrights on how to do AI training. As AI technology advances, who they are being trained on and how they’re being trained increasingly becomes an issue. The settlement of these legal cases could have far-reaching impacts in the area of AI, especially in making companies adopt more open and legal methods of acquiring training data.