Official statements or press releases from Meta

Meta Faces California Lawsuit Over LibGen Copyright Use in AI Training

Jan 9, 2025, 09:04 PM

Meta Platforms, under the approval of CEO Mark Zuckerberg, is facing allegations of copyright infringement in a lawsuit filed by authors including Sarah Silverman and Ta-Nehisi Coates. Newly unredacted court documents reveal that Meta used a dataset from LibGen, a shadow library containing pirated e-books and articles, to train its Llama AI models. Internal documents indicate that Meta employees were aware of the dataset's pirated nature, with concerns raised about its potential legal and regulatory implications. The lawsuit, filed in a California court, accuses Meta of violating the California Comprehensive Computer Data Access and Fraud Act (CDAFA), stripping copyright information from the materials to conceal infringement, and engaging in torrenting the LibGen dataset, further distributing its contents. Ahmad Al-Dahle, Meta's head of generative AI, is alleged to have approved the torrenting despite internal reservations. The case, Kadrey v. Meta, highlights broader concerns about the use of copyrighted material in AI training and the applicability of the fair use doctrine. Meta has argued that its actions are protected under fair use, but the allegations have raised significant ethical and legal questions about its practices.