AI Copyright Lawsuit Targets Meta Over Book and Journal Training Data
Major publishers and authors filed a class action alleging Meta used copyrighted works to train AI.
Why it matters: The suit highlights escalating fair use litigation risks for AI companies and introduces new challenges to standard AI training practices. A class-action outcome could reshape copyright liability for large language models and their developers.
- Five publishing giants and author Scott Turow filed a class-action against Meta and Mark Zuckerberg on May 5, 2026.
- The lawsuit claims Meta reproduced and distributed millions of books and articles without permission or payment.
- Meta will contest the suit, arguing AI training on copyrighted content can be fair use.
- In 2025, Anthropic settled a similar class action for $1.5 billion after training AI on copyrighted books.
On May 5, 2026, Elsevier, Cengage, Hachette Book Group, Macmillan, McGraw Hill, and author Scott Turow filed a class-action lawsuit targeting Meta and CEO Mark Zuckerberg. The complaint alleges Meta used millions of copyrighted books and journal articles as training data for its Llama AI platform without consent or compensation to rights holders.
The plaintiffs called Meta's conduct “one of the most massive infringements of copyrighted material in history,” accusing it of reproducing and distributing protected works at unprecedented scale. Their suit seeks to expand protection for authors whose work powers modern AI systems, with high-profile authors like James Patterson and Donna Tartt named among those affected. (Washington Post)
Meta contends it will “fight this lawsuit aggressively,” maintaining that training AI on copyrighted material falls under fair use when done in accordance with legal standards. (Washington Post)
These claims follow a 2025 case in which Anthropic agreed to pay $1.5 billion to settle a class action initiated by authors, after allegedly using millions of copyrighted works to train its models. In that dispute, a federal judge ultimately ruled that AI training may constitute fair use—provided books are legally obtained—but allowed litigation to proceed if pirated content was involved. (Washington Post - Anthropic settlement; Washington Post - fair use ruling)
- The Meta class action could set a new precedent for how US courts address the intersection of copyright and AI.
- AI developers and in-house legal teams must watch how courts handle the scale of alleged infringements and evolving fair use defenses.
By the numbers:
- $1.5B — Anthropic’s 2025 settlement for alleged AI copyright infringement
- 5 — Major publishing houses suing Meta
- Millions — Alleged copyrighted works used by Meta to train Llama
Yes, but: A 2025 federal court found AI training on copyrighted works could be fair use if the material was legally acquired—potentially narrowing liability for some developers.