Book publishers sue Google for copyright infringement over Gemini AI training | books

A group of major publishers has filed a lawsuit against Google, accusing the company of illegally using millions of copyrighted books to help build Gemini AI models, in “one of the most widespread infringements of copyrighted material in history.”

The case, filed in federal court in New York, was brought by three publishers – Hachette Book Group, Cengage Learning, and Elsevier – and best-selling American author Scott Turow.

Publishers argue that Google has repurposed books that were made available to limited services such as Google Books, Google Play Books, and Google Scholar. These services allowed Google to use the works in specific ways — for example, to display searchable snippets or sell e-books — but not to copy them to train commercial AI products, the lawsuit alleges.

“In a desperate attempt to maintain its dominance on the Internet, Google abandoned its early slogan ‘Don’t be evil’ and engaged in one of the most widespread infringements of copyrighted material in history,” the lawsuit said.

According to the complaint, the technology company made copies of copyrighted Gemini training books without obtaining permission or payment, despite internal discussions recognizing the legal risks. The filing claims that Google internally reported that it could face “potential fines between $10 billion and $100 billion” for using text provided by publishers for Google Play Books.

Publishers say Google’s actions harm authors and the publishing industry more broadly, arguing that AI-generated content could negatively impact book sales.

The document notes that, for example, Gemini can create “a 100-page murder mystery set in a quiet seaside town full of secrets, which replaces an original copyrighted murder mystery that Gemini rehearsed” in 20 minutes for 39 cents. “No publisher or author can compete with that.”

The lawsuit names a number of specific books that the publishers claim are among the copyrighted works used without permission, including N. K. Jemisin’s The Fifth Season, Lemony Snicket’s Who Can That Be at This Hour?

This case adds to the growing legal battle over artificial intelligence and copyright. Authors and publishers have filed a series of lawsuits against Google, OpenAI, Anthropic, and Meta, alleging that their copyrighted works have been used without permission to train AI models. These include copyright lawsuits brought by a group of authors in which a judge ruled in favor of Meta last June, and a landmark settlement in which Anthropic agreed to pay $1.5 billion to authors who claimed that pirated copies of their books were used to train the chatbot Cloud.

Earlier this year, thousands of authors, including Kazuo Ishiguro, Philippa Gregory, and Richard Osman, published “Blank” to protest AI companies using their work without permission.

The new case follows an earlier attempt by Hachette and Cengage to join an existing copyright lawsuit filed against Google by authors and illustrators in 2023. Google has opposed their participation in that case, prompting the publishers to launch a separate action.

The plaintiffs are seeking statutory damages, a permanent injunction barring Google from continuing the alleged infringement, and a court order requiring the company to destroy any unauthorized copies of their work used to train its AI systems.

Google did not respond to the Guardian’s request for comment.