Publishers sue Google over alleged use of copyrighted works to train Gemini

On July 10, Hachette Book Group, Cengage Learning, Elsevier and author Scott Turow filed a putative class action against Google in federal court in New York, accusing the company of willfully infringing millions of textual works to build and train its Gemini large language models. The suit covers a proposed class of authors and publishers and reaches fiction, nonfiction, children’s books, memoirs, poetry, educational works and scholarly articles across thousands of subject areas.

Google drew on books and journal articles obtained through Google Books and other Google services, along with unauthorized web scrapes from across the internet, including pirate sources and material behind paywalls. It copied those works repeatedly for AI training and stripped copyright management information to hide where the training data came from. The publishers also allege Google used content acquired for limited purposes, such as snippet display or ebook distribution, for a separate commercial training project the original agreements did not authorize.

Hachette and Cengage previously moved to intervene in In re Google Generative AI Copyright Litigation, a case pending before Judge Eumi K. Lee in the Northern District of California. The publishers brought a separate action to preserve claims outside the class action already underway. Google internally flagged use of publisher-provided copyrighted books for AI as highly problematic and warned of potential fines ranging from $10 billion to $100 billion, while also identifying risks that publishers would treat LLM training on their books as infringement and pull their content or sue.

Anthropic has agreed to a $1.5 billion settlement with authors in a separate book-training class action, while The New York Times’ copyright case against OpenAI and Microsoft remains active in Manhattan federal court.