Book Publishers File Class Action Against Meta Over Llama AI Training
Key Points
- Five major publishers—Hachette, Macmillan, McGraw Hill, Elsevier, Cengage—and author Scott Turow have filed a class‑action lawsuit against Meta.
- The suit alleges Meta used millions of copyrighted works to train its Llama AI model without permission or compensation.
- Complaint claims Mark Zuckerberg personally authorized the alleged infringement.
- Meta argues that using copyrighted material for AI training falls under fair use, citing recent court opinions.
- Previous lawsuits against Meta over AI training have been dismissed or remain unresolved.
- A judge in a related Anthropic case suggested piracy could be a viable claim for authors.
- Publishers seek statutory damages, actual damages, and an injunction to stop further use of their works.
- The case could set precedent for how tech companies license content for AI development.
Five major book publishers and author Scott Turow have sued Meta and CEO Mark Zuckerberg, accusing the company of illegally using millions of copyrighted works to train its Llama generative‑AI model. The complaint alleges that Meta reproduced and distributed the material without permission or compensation, and that Zuckerberg personally authorized the infringement. Meta has faced multiple lawsuits over its AI training practices, but the company maintains that such use falls under fair‑use doctrine.
Hachette, Macmillan, McGraw Hill, Elsevier, Cengage and bestselling author Scott Turow have united in a class‑action lawsuit that targets Meta Platforms Inc. and its chief executive, Mark Zuckerberg. The plaintiffs contend that Meta fed millions of copyrighted books and articles into its Llama generative‑AI system without securing permission or paying royalties.
According to the filing, Meta’s actions "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." The complaint goes further, asserting that Zuckerberg himself "personally authorized and actively encouraged the infringement."
The suit adds to a growing roster of legal challenges aimed at the tech giant’s AI endeavors. Earlier this year, a different coalition of authors brought a copyright case against Meta that ultimately failed. In a separate matter, a lawsuit filed by the LibGen organization singled out Zuckerberg for encouraging the use of protected works. Last year, a group of British authors also raised concerns about Meta’s training practices, though that effort has not yet reached court.
Meta’s defense hinges on the argument that training AI models on copyrighted material can qualify as fair use. Company spokesperson Dave Arnold told The New York Times that courts have "rightly found that training AI on copyrighted material can qualify as fair use," emphasizing the transformative potential of AI for productivity and creativity.
Judicial precedent offers mixed signals. In a related case against Anthropic, the rival AI firm, a judge was not persuaded by the copyright infringement claim but suggested piracy could be an alternative pathway for authors seeking damages. The Anthropic ruling underscores the legal uncertainty surrounding AI training data and the standards courts will apply.
If the plaintiffs succeed, the lawsuit could reshape how tech firms acquire and use copyrighted content for AI development. Publishers argue that the current practice deprives them of revenue and undermines the value of their intellectual property. They are seeking statutory damages, actual damages, and an injunction to halt further use of their works in Meta’s training pipelines.
Meta has faced criticism for its opaque data‑collection methods, especially as AI models become more central to its product strategy. The company’s Llama series, positioned as an open‑source alternative to other large language models, relies heavily on vast datasets to achieve its capabilities. Critics argue that the scale of data required inevitably pulls in copyrighted material unless rigorous licensing mechanisms are put in place.
The outcome of this case may influence industry standards for AI training data, prompting tech firms to negotiate clearer licensing agreements with content creators. Until a court renders a decision, the debate over the balance between innovation and intellectual‑property rights is likely to intensify.