Publishers take Google to court over Gemini training

A new copyright fight is taking aim at Google and its Gemini models. Hachette Book Group, Cengage Learning, Elsevier, and author Scott Turow are seeking a class action lawsuit, alleging the company used their works without permission or payment to train its AI systems.

According to the complaint, Google “reproduced millions of copyrighted works without permission, without providing any compensation to authors or publishers, and with full knowledge that its conduct violated copyright law.” The filing also claims Google removed CMI — copyright management information — from the works it allegedly used, in order to hide training sources and enable unauthorized use.

The lawsuit does not stop at training data. It also argues that Gemini can produce and, at times, encourage the creation of derivative or “copycat” works without attribution or compensation. The complaint says Google knows Gemini could keep generating outputs that act as substitutes for the copyrighted works it was trained on, and alleges the company has failed to put effective guardrails in place.

Recommended reading

DeepSeek eyes $1.5B raise at $71B ahead of IPO

The case adds to a growing wave of legal pressure from publishers and authors seeking compensation from AI companies over training materials. Several of the same parties are already involved in a similar class action against Meta.

So far, though, copyright claims against AI companies have had mixed results. In 2025, another group of writers reached an initial $1.5 billion settlement with Anthropic in a case tied to alleged copyright infringement and piracy involving Claude, but the judge rejected it as “nowhere near complete.” Separate efforts by authors against Meta fell short last year, while another pair of authors has also challenged Apple over alleged unlicensed use of their work for AI training.