French publishers take Meta to court over AI copyright infringement

French publishers and authors have filed a lawsuit against Meta Platforms Inc., accusing the technology company of copyright infringement for allegedly using their books to train its generative artificial intelligence model without permission.

According to Bloomberg, the complaint was lodged this week in a Paris court specialising in intellectual property by the Syndicat National de l'Édition (SNE), which represents major French publishers such as Hachette and Editis, along with authors' association Société des Gens de Lettres (SGDL) and writers' union Syndicat National des Auteurs et des Compositeurs (SNAC). The legal action was announced at a press conference on Wednesday.

SNE president Vincent Montagne stated that the group had collected evidence of "massive" copyright breaches and had previously reached out to Meta without success. He added that the European Commission had been informed, asserting that Meta's practices violate EU regulations on artificial intelligence.

Generative AI language models, such as Meta's Llama and OpenAI's ChatGPT, rely on extensive text data, including books and articles, for training. This has sparked a series of lawsuits worldwide, with content creators arguing that using their work to train AI models constitutes copyright infringement. AI companies have generally been reluctant to disclose their data sources, maintaining that the practice falls under "fair use" as defined by US copyright law.

Similar legal battles have emerged globally. In December 2023, The New York Times sued OpenAI and Microsoft Corp. over the use of its articles to train large-language models. In April 2024, a group of authors filed a class-action lawsuit against Anthropic, backed by Amazon.com Inc., for allegedly using their books to train its AI system. Indian publishers also launched a legal challenge against OpenAI in January. British musicians in the UK released a silent album to protest proposed AI law changes, fearing a loss of creative control.

The case in France adds to the growing legal and regulatory scrutiny surrounding AI training practices, raising questions about intellectual property rights and the boundaries of AI development.