Five major publishers and Scott Turow sue Meta and Zuckerberg over pirated books used to train Llama

What happened

On 5 May 2026, five major publishing houses — Elsevier, Cengage, Hachette Book Group, Macmillan and McGraw Hill — and bestselling author Scott Turow filed a putative class action against Meta Platforms and its founder and CEO Mark Zuckerberg in the U.S. District Court for the Southern District of New York (Case No. 1:26-cv-03689), alleging willful copyright infringement of millions of textual works, including literature, educational works and scholarly articles, to develop Meta's Llama large language models.

What the complaint alleges

The plaintiffs allege Meta copied millions of copyrighted works without permission, first by torrenting books and journal articles from pirate sites including LibGen and Anna's Archive and by downloading unauthorized web scrapes of virtually the entire internet, including subscription-only content, and then by copying those works repeatedly in the course of training Llama. The complaint alleges the decisions to infringe were deliberate and made at Zuckerberg's direction, and that Meta chose infringement over licensing even while negotiating dataset licenses in early 2023. The plaintiffs seek monetary damages and injunctive relief, including an order requiring Meta to destroy all infringing copies in its possession and a full disclosure of the copyrighted works used to train Llama.

Meta's position

Meta has denied wrongdoing, stating that AI training on copyrighted material can qualify as fair use under U.S. law.

Context

The suit targets how Meta obtained training material rather than training itself. It follows Kadrey v. Meta, in which a California federal court granted Meta summary judgment in June 2025 on the training question, and Bartz v. Anthropic, in which Anthropic agreed to a $1.5 billion settlement over pirated books, granted final approval in July 2026. Both of those cases are already covered on this site; this publishers' suit against Meta was not.