Court documents reveal that Zuckerberg approved use of protected works, according to US lawsuit
Meta used pirated versions of copyrighted books to train its artificial intelligence (AI) systems with approval from CEO Mark Zuckerberg, authors claim in court documents released on Wednesday (10.Jan.2025) in the United States.
The lawsuit, filed in 2023 in California federal court, includes writer Ta-Nehisi Coates and comedian Sarah Silverman among its plaintiffs. They accuse the parent company of Facebook and Instagram of improperly using their works to train the LLaMA language model.
According to , internal documents reveal that Zuckerberg approved the use of the LibGen dataset, which contains millions of pirated works distributed via peer-to-peer torrents, even after warnings from the executive team. An internal communication cited in the lawsuit states that Meta used “a set of data that we know is pirated”.
District Judge Vince Chhabria in 2024 rejected initial claims that text generated by Meta’s chatbots infringed copyright. On Thursday (Jan 11, 2025), he allowed the plaintiffs to file an updated complaint, but expressed skepticism about the new arguments.
The case is part of a series of lawsuits against technology companies over their use of copyrighted materials in the development of AI. The companies argue that their practices fall within the concept of “fair use”, which allows limited use of protected works in certain circumstances.
