Publishers and Authors Sue Google Over Gemini AI Training

Google is facing another legal challenge over its artificial intelligence practices after a coalition of publishers and authors filed a class action lawsuit alleging that the company used copyrighted books without permission to train its Gemini AI models. The lawsuit argues that Google copied protected works for AI development and, in some cases, altered or removed copyright management information to hide that the materials had been used during training.

The plaintiffs include the publishers Hachette, Cengage, Elsevier, and S.C.R.I.B.E., as well as author Scott Turow. According to the complaint, Google’s actions exceeded the permissions granted by copyright holders and allowed the company to use protected works for purposes unrelated to the original agreements.

"Google illegally copied works from all these scope-limited programs for AI training, knowing it lacked authorization to do so," the lawsuit states.

The case adds to a growing wave of lawsuits filed by publishers, authors, artists, and other copyright owners against major AI developers, including Google, Meta, OpenAI, and Anthropic. These cases centre on whether technology companies can legally use copyrighted material to train generative AI systems without obtaining permission from rights holders.

Several of those disputes remain unresolved, although two recent federal court decisions in California favoured AI companies by finding that using copyrighted works for AI training qualified as fair use under existing U.S. copyright law. Those rulings relied on legislation written long before the internet and generative AI emerged, leaving broader legal questions about AI training far from settled.

The source material also notes that Anthropic was ordered to pay $1.5 billion for pirating works used to train its AI models, marking the largest payout in U.S. copyright law history. Around 500,000 writers became eligible for payments of at least $3,000, although many declined the settlement to continue pursuing separate legal claims over AI training.

Unlike the California cases, the lawsuit against Google was filed in the U.S. District Court for the Southern District of New York. Because a different federal judge will consider it, the case could provide another interpretation of whether Google’s use of copyrighted books qualifies as fair use.

The complaint argues that Google’s relationship with publishers is distinct from many other AI copyright disputes. For years, publishers and authors voluntarily supplied books through Google Books so users could search titles, discover publications, and view limited snippets together with bibliographic information. According to the lawsuit, those arrangements were designed to improve search functionality rather than provide training material for commercial AI systems.

The plaintiffs allege that Google later trained Gemini using copies of books obtained through Google Books and titles uploaded to the Google Play Store without securing additional authorisation from copyright owners. They argue that the company exceeded the limited purposes for which those books were originally provided.

The lawsuit also references an internal Google document that allegedly warned employees that relying on copyrighted books for AI training could be “highly problematic for Google” and expose the company to potential fines ranging from tens of billions to hundreds of billions of dollars.

The latest lawsuit represents another closely watched test of how copyright law will be applied to generative AI as courts continue weighing the rights of creators against the rapidly expanding use of copyrighted material in AI model development.