Overview
- The class action, filed Tuesday in Manhattan, was brought by Macmillan, Hachette, McGraw Hill, Cengage, Elsevier and author Scott Turow.
- Plaintiffs say Meta staff pulled millions of books and articles from shadow libraries such as OceanofPDF, stripped copyright notices, and used the files to train Llama, alleging Mark Zuckerberg approved the plan.
- The suit seeks a court order forcing Meta to destroy the copies and stop using them, arguing AI-made summaries and imitations could replace originals and undercut writers’ incomes.
- As evidence, the filing cites Llama’s own statements that it was trained on works by specific authors, along with earlier Meta acknowledgments of downloading data from shadow libraries via BitTorrent.
- Meta says it will fight the case and points to fair use rulings, while mixed outcomes in other AI cases and Anthropic’s $1.5 billion settlement highlight the legal and financial stakes for both tech firms and publishers.