A coalition of publishers and an author filed a class action lawsuit against Google in a New York federal court this week, accusing the company of copying millions of copyrighted books without authorisation to train its Gemini artificial intelligence model. The plaintiffs — Hachette Book Group, Cengage Learning, Elsevier, author Scott Turow and his publishing company S.C.R.I.B.E. — allege that Google secretly reproduced works originally provided to its Google Books service and other platforms for limited purposes, then used that material to develop a commercial AI system that now generates content directly competing with the original authors.

According to the court filing, the plaintiffs claim that Gemini can produce books at a scale and speed they describe as unprecedented, and that the model tailors outputs to mimic the expressive elements and creative choices of specific writers. The lawsuit further asserts that Google removed copyright management information from the works to conceal their origin. The plaintiffs are seeking an injunction to halt the alleged infringement and unspecified monetary damages. Google has not yet issued a public response to the filing, though the complaint references internal company documents that reportedly assessed the legal exposure from such training as potentially reaching tens of billions of dollars.

The case is the latest in a series of copyright challenges confronting AI developers in the United States. The same group of plaintiffs previously sued Meta on similar grounds; in that litigation, a federal judge in California ruled that the use of copyrighted materials to train AI models could qualify as fair use under US law, handing Meta a partial victory. Separately, Anthropic agreed to a $1.5 billion settlement with authors who accused it of illegally copying books to train its Claude model, though a judge in that matter also found that the training itself could be considered transformative fair use while deeming other uses of pirated materials unlawful. Legal observers in Washington and Silicon Valley note that these mixed rulings leave the boundaries of fair use in AI training unsettled, with the Google case poised to test the doctrine further.

The lawsuit, filed in the Southern District of New York, requests class action status on behalf of other rights holders. No hearing date has been set. The plaintiffs are asking the court to order the destruction of unauthorised copies of their works and to award statutory damages. The case will now proceed through initial pleadings, with Google expected to respond in the coming weeks. The outcome is being watched closely by publishing and technology sectors on both sides of the Atlantic, as it may influence how AI companies negotiate access to copyrighted training data.