Hachette, Elsevier, Cengage and Scott Turow Sue Google Over Gemini AI Training

<img src="/wp-content/uploads/2026/02/ai_authored_book_legalities_explained_0dk0t.jpg" alt="Book publishers and an author file a copyright lawsuit against Google over Gemini AI training" style="width:100%;border-radius:12px;margin-bottom:2rem;">

Hachette Book Group, Cengage Learning and Elsevier, together with bestselling author Scott Turow, filed a putative class action lawsuit against Google on July 13, 2026, accusing the company of using millions of copyrighted books and journal articles to train its Gemini AI models without permission. According to reporting by TechCrunch and a statement from Hachette Book Group, the complaint was filed in the U.S. District Court for the Southern District of New York and marks one of the largest publisher-led challenges yet to how a major AI lab sourced its training data.

What the publishers are claiming

The plaintiffs allege that Google trained early Gemini models on books and texts pulled from Google Books, Google Play and Google Scholar, using them far outside the limited purposes those programs were originally built for, according to Publishers Weekly and Good e-Reader. Hachette's own statement describes the conduct as "willful copyright infringement." Multiple outlets, including TheWrap and Gizmodo, report that the complaint also accuses Google of removing or altering copyright management information on the works to help conceal that Gemini was trained on unauthorized material.

The lawsuit points to Google's own internal assessment of the risk. According to reporting on the case, an internal document is said to describe using copyrighted books for AI training as "highly problematic for Google," warning it could expose the company to "$10Bs-$100Bs in potential fines."

The complaint frames the stakes in stark terms for the industry, arguing that if the conduct goes unaddressed, Google "will continue to infringe Plaintiffs' and the Class's rights, cause broad and lasting damage to the literary industry and authors, and weaken the incentive to create that is at the core of the Copyright Act." According to amNewYork's coverage, the plaintiffs also argue that Gemini now competes directly with the books it learned from by generating detailed summaries and textbook-style explanations that could reduce demand for the original works.

What the publishers want from the court

The plaintiffs are asking the court for statutory damages, a permanent injunction to stop further infringement, and an order requiring Google to destroy any unauthorized copies of their works, according to reporting on the filing. As of publication, Google had not issued a public statement responding to the specific allegations in this complaint.

Why this matters for authors

This is not the first major publisher lawsuit against an AI company this year. Hachette, Macmillan, McGraw Hill, Elsevier and Cengage joined Scott Turow in a similar suit against Meta in May 2026 over its Llama models, and authors separately sued Anthropic over Claude's training data. Each case is testing a slightly different question: not just whether training AI on books can be fair use, but whether the way a company obtained those books, through scraped platforms, pirated sites, or repurposed programs like Google Books, changes the legal answer.

For authors who write with AI tools, the practical takeaway has not changed. The legal fight over training data is between publishers, authors' groups and the AI companies themselves, and it will likely take years to resolve. What you can control is how you use these tools in your own work: draft with AI, then revise in your own voice, and keep records of how a manuscript came together. Our guide on how to edit and humanize AI-generated book content walks through that process in more detail.

PageWriter Studio is built around that same principle: AI drafts the outline and chapters, and you shape every revision with a style profile tuned to your own voice, so the manuscript stays yours regardless of how the courts eventually rule on any single model's training data. If you want to try that workflow, you can start a free trial and keep full control of your next book from the first page.