Three major publishers and a novelist say Google turned a library program into an AI pipeline
Hachette Book Group, Cengage Learning, and Elsevier uploaded their books to Google Books so readers could search them. According to a lawsuit filed in New York federal court, Google took those same books and fed them into Gemini, its commercial AI product, without asking permission or offering payment.
Author Scott Turow joined the three publishers in filing the suit, which requests class-action status. The complaint describes the alleged conduct as "one of the most prolific infringements of copyrighted materials in history."
##A library program repurposed
The core legal argument turns on scope. Google Books was established under agreements that granted limited rights: digitization for search and snippet display. The plaintiffs argue Google exceeded those permissions by using the same digitized texts as training data for a commercial AI system, a purpose never authorized under the original arrangements.
The complaint also alleges Google scraped copyrighted works from the web beyond what Google Books contained, compounding the unauthorized copying.
Get the latest model rankings, product launches, and evaluation insights delivered to your inbox.
The plaintiffs say Google stripped copyright management information from the works it ingested, the embedded metadata that identifies authors, rights holders, and licensing terms. Under U.S. copyright law, removing that information is itself a violation, and the complaint frames it as an attempt to obscure the training data's origins.
##Internal warnings
The lawsuit alleges Google proceeded despite internal warnings about the legal risks of using copyrighted material for AI training. The complaint does not detail who raised those warnings or when, and that account comes from the plaintiffs rather than independent documentation. If the allegation holds up in discovery, it would undercut any argument that Google acted in good faith.
Google has not issued a public response to the suit.
##Where this fits
AI copyright litigation has accumulated across multiple companies and content categories since 2023, with disputes involving news publishers, visual artists, and now book publishers pressing similar theories about training data. The Gemini suit stands out for the scale of the plaintiffs. Elsevier alone publishes a substantial share of the world's scientific journals. The plaintiffs also assert that a pre-existing content partnership was converted into an AI data pipeline without renegotiation.
The class-action structure, if certified, would extend the case beyond the named plaintiffs to thousands of authors and publishers whose works appeared in Google Books or were scraped from the web.
The case is pending in the Southern District of New York. A scheduling order setting discovery timelines and any early motion practice will be the next marker of how quickly this proceeds toward trial.