Google sued over copyright infringement claims in Gemini AI training

Google has been sued in the United States by several major publishers and an author for allegedly using copyrighted books without permission to train its Gemini artificial intelligence model. The lawsuit was filed in a federal court in New York on Tuesday, July 14.

The plaintiffs include publishing houses Hachette Book Group, Cengage Learning and Elsevier, along with American novelist Scott Turow.

The lawsuit alleges that Google used books obtained from publishers for Google Books, Google Play Books and Google Scholar beyond the authorised scope to train its Gemini AI model. The plaintiffs argue that while permission was granted to make book content searchable or for e-book sales, it was not authorised for commercial AI model development.

The complaint further claims that Google created copies of millions of copyrighted books without permission or compensation to train Gemini. It also alleges that internal company documents acknowledged the legal risks of such use.

According to the publishers, AI trained in this manner can generate text resembling new books in a very short time and at low cost, potentially harming authors' and publishers' income and book sales. The complaint cites an example where Gemini can generate a 100-page mystery novel in just 20 minutes for about 39 cents.

The lawsuit lists several books, including N.K. Jemisin's 'The Fifth Season' and Lemony Snicket's 'Who Could That Be at This Hour?', as examples of works used without authorisation.

Google, along with OpenAI, Anthropic and Meta, has faced multiple lawsuits over the use of copyrighted books in AI training. This new lawsuit is part of the ongoing legal battle.

The plaintiffs have sought compensation, a permanent injunction to prevent future use, and an order to destroy all unauthorised copies of books used for AI training. Google has yet to make any official comment on the matter.