Publishers sue Google over alleged use of books to train Gemini AI

Multiple book publishers sued Google on Tuesday, accusing the company of copying copyrighted content to train artificial intelligence models and then using the technology to create work that competes with the original authors.

The lawsuit asks for class action status and was filed in New York by Hachette Book Group, Cengage Learning, Elsevier, author Scott Turow and his publishing company S.C.R.I.B.E. The plaintiffs allege that Google secretly copied millions of works that had been provided to Google Books and other services for limited purposes, and then used those works to train Gemini, Google's AI model.

The lawsuit says Gemini can create books at a scale and speed that competes with human writers. It also alleges that Gemini can tailor outputs to mimic the expressive elements and creative choices of specific authors.

The plaintiffs asked the court for an injunction and an unspecified amount of damages.

The case is part of a series of copyright lawsuits against AI developers. In May, several publishers, including Hachette, Cengage, Elsevier and Turow, sued Meta in New York on similar grounds.

A US judge in September approved a $1.5 billion settlement between Anthropic and several authors who claimed the company illegally copied their work to train its AI model, Claude. In that case, Anthropic won a partial victory when a judge ruled that using books to train Claude was transformative enough to qualify as fair use under US law, while other uses of pirated materials did not qualify.

Meta also won a partial victory last year when a US judge in San Francisco ruled that its use of copyrighted materials was fair use. That case was filed by comedian Sarah Silverman, author Ta-Nehisi Coates and others.