Five of the world’s largest publishing houses have launched a class-action lawsuit against Meta Platforms in a Manhattan federal court, accusing the Mark Zuckerberg-led tech giant of pirating millions of copyrighted works to train its Llama artificial intelligence models.
Elsevier, Cengage, Hachette, Macmillan and McGraw Hill, joined by bestselling American author Scott Turow, filed proceedings on Tuesday alleging that Meta knowingly used pirated copies of textbooks, peer-reviewed scientific journals and novels—including N.K. Jemisin’s The Fifth Season and Peter Brown’s The Wild Robot—to train the systems that underpin Meta’s generative AI products.
The complaint, which seeks unspecified damages and class-action status on behalf of a wider pool of rights holders, marks the first time that academic and trade publishers have moved against Meta as a unified front.
Maria Pallante, president of the Association of American Publishers, said: “Meta’s mass-scale infringement isn’t public progress, and AI will never be properly realised if tech companies prioritise pirate sites over scholarship and imagination.”
Meta has signalled it will mount a robust defence. “AI is powering transformative innovations, productivity and creativity for individuals and companies, and courts have rightly found that training AI on copyrighted material can qualify as fair use,” a spokesperson said. “We will fight this lawsuit aggressively.”
The case opens yet another front in a war that is rapidly redrawing the commercial map for content owners. Dozens of plaintiffs, from The New York Times pursuing OpenAI and Microsoft, to a coalition of authors, news outlets and visual artists, have already filed suit against leading AI developers.
The first major scalp came when Anthropic, the AI company backed by Amazon and Google, agreed in 2025 to pay $1.5 billion to settle a class action brought by a group of authors.
Benjamin Woollams, chief executive of TrueRights, argues the sector urgently needs commercial infrastructure capable of matching the speed at which AI models are being built. “Every one of these lawsuits points to the same underlying problem: there’s no standardised way to license creative work and likeness for AI,” he said.
“Tech companies aren’t villains for wanting training data, and creators aren’t luddites for wanting to be paid, but the infrastructure to connect them simply hasn’t existed until now. This represents a huge opportunity for those in the industry to build a transparent and trusted licensing framework that allows innovation and creator rights to coexist commercially.”
For Meta, the stakes extend beyond the immediate price tag. A successful class certification could expose the group to claims from thousands of rights holders, while an adverse ruling would reverberate across an industry that has built its competitive edge on the unrestricted ingestion of human-authored work.