Full Breakdown
Publishers and Author Sue Meta Over Alleged Copyright Infringement in Llama AI Training
5/8/2026, 3:47:56 AM
Alleged Unauthorized Use of Copyrighted Works
On May 5 2026, five publishers—Hachette, Macmillan, McGraw Hill, Elsevier, and Cengage—joined novelist Scott Turow in a class-action filed in the Southern District of New York against Meta Platforms and CEO Mark Zuckerberg. The complaint says Meta used millions of copyrighted books and journal articles, obtained from pirate sites such as LibGen and Sci-Hub, to train its Llama language model, removing copyright-management information. Internal emails reveal a licensing-budget discussion that grew from $17 million to $200 million in early 2023 but was halted after Zuckerberg’s approval. The suit quantifies the alleged theft as over 267 terabytes of material—equivalent to hundreds of millions of publications and many times the Library of Congress’s print collection. Plaintiffs argue Llama can produce verbatim excerpts, replacement chapters, and summaries that substitute for the original works.
Background & Prior Litigation
The case follows a series of copyright actions involving generative-AI. In 2025, Anthropic settled a class-action by authors for $1.5 billion. A 2023 suit by 13 authors against Meta was dismissed after a judge ruled that training data qualified as fair use. Meta has repeatedly cited those rulings in public statements.
Key Plaintiffs and Defendants
- Plaintiffs: Hachette, Macmillan, McGraw Hill, Elsevier, Cengage, and author Scott Turow.
- Defendants: Meta Platforms, Inc.; Mark Zuckerberg.
- AI system: Llama.
Official Statements from Meta
Meta spokesperson Dave Arnold said the company sees AI as a productivity driver, notes that courts have recognized fair-use for AI training data, and said Meta will contest the suit vigorously.
Publisher and Author Criticism
Macmillan CEO Jon Yaged called the conduct “unconscionable” and stressed the Copyright Act’s role in protecting creators. McGraw Hill CEO Philip Moyer urged protection of authors’ rights and noted a market for licensing. AAP CEO Maria Pallante said Meta chose to “steal” literary works.
Verbatim Quotes
- “Defendants reproduced and distributed millions of copyrighted works without permission, without providing any compensation to authors or publishers, and with full knowledge that their conduct violated copyright law,” — Plaintiffs’ complaint
- “Zuckerberg himself personally authorized and actively encouraged the infringement.” — Plaintiffs’ complaint
- “if we license once [sic] single book, we won’t be able to lean into the fair use strategy.” — Meta employee (cited by Publishers Weekly)
- “The Copyright Act has long been the foundation for safeguarding intellectual property. That protection is needed now more than ever,” — Jon Yaged, CEO, Macmillan
- “We will fight this lawsuit aggressively.” — Dave Arnold, Meta spokesperson
Potential Impact and Next Steps
The plaintiffs seek monetary damages and an injunction to delete any infringing copies held by Meta. A ruling limiting fair-use could force AI firms to license training data, reshaping industry practice. Both sides have signaled readiness for prolonged litigation, and the decision will likely set precedent for future AI copyright disputes.
