Particle.news
Download on the App Store

Publishers and Scott Turow Sue Google Over Gemini’s Use of Copyrighted Books

The lawsuit could force courts to define when using copyrighted texts to train large language models is lawful.

Overview

  • The complaint, filed in New York on Tuesday July 14, says Hachette, Cengage, Elsevier and author Scott Turow accused Google of copying millions of works to train its Gemini model.
  • Plaintiffs allege Google took books from Google Books and other limited‑use services without permission and ask the court to order the practice to stop and to award unspecified damages.
  • The suit says Gemini can quickly produce book‑length works that compete with human authors and that the model can mimic distinctive authors’ stylistic and creative choices.
  • The case joins a wave of publisher lawsuits over AI training data and follows an Anthropic settlement for at least $1.5 billion and a judge’s earlier ruling that merely feeding protected works into a model is not automatically infringement.
  • A ruling for the publishers could force AI companies to change how they assemble training corpora, push firms toward licensing deals, and affect authors’ incomes by limiting model outputs that replicate their work.