Zuckerberg 'personally approved and encouraged' Meta's copyright infringement

Five publishers (Hachette, Macmillan, McGraw Hill, Elsevier, Cengage) and author Scott Turow filed a lawsuit against Meta and CEO Mark Zuckerberg in the Southern District of New York federal court on May 5.

The plaintiffs claim that Meta illegally copied millions of books, academic journals, and articles and used them to train generative AI systems such as Llama. The complaint states that Meta torrented materials from pirate sites and repeatedly reproduced scraped data covering nearly the entire web.

The lawsuit also alleges that while Meta once considered large-scale licensing, it abandoned the licensing strategy on Zuckerberg's instructions. Internal memos reportedly acknowledged LibGen as a "pirated" dataset while stating an intent not to disclose that it had been used for training.

The plaintiffs allege that Meta removed copyright management information (CMI) to hide the training sources and torrented as much as 267TB of illegal material. Meta, on the other hand, maintains that AI training constitutes fair use and has stated it will vigorously defend against the lawsuit.

This case is not merely a dispute over damages—it could affect Meta's entire Llama strategy regarding the legality of training data for large models and whether licensing was being evaded.