Publishers Sue Google Over Alleged Piracy of Books Used to Train AI - Sarmad

A group of publishing houses and authors filed lawsuits against Google, accusing the tech giant of copyright infringement by using protected content to train its artificial intelligence models, and then generating content that directly competes with the original works of the authors.

The lawsuit states that “the scale and speed of Gemini’s ability to produce books and compete with human writers is unprecedented.”

The case was filed in a New York court as a class-action lawsuit on behalf of publishing houses Hachette Book Group, McGraw Hill Learning, and Elsevier, as well as author Scott Turow and his publishing company S.C.R.I.B.E.

The plaintiffs accuse Google of “secretly copying millions of works” that it had obtained through Google Books and other services for specific purposes, and then using those materials to train its Gemini AI model.

They further assert that the content generated by Gemini directly competes with the original works authored by the rights holders.

The lawsuit adds that “Gemini even tailors its outputs to simulate the expressive elements and creative choices of specific authors.”

This is the latest copyright lawsuit filed against artificial intelligence development companies.

In May, a group of publishers including Hachette, McGraw Hill, and Elsevier, along with Scott Turow, filed a similar lawsuit against Meta in a New York court.

In September, a U.S. judge approved a $1.5 billion settlement between Anthropic and several authors who accused the company of illegally copying their works to train its “Claude” AI model.

The decision was seen as a partial victory for Anthropic, as the judge ruled that using books to train the model could be considered “fair use” under U.S. law, while deeming other uses of pirated materials illegal.