Publishers Sue Google Over Gemini Training on Millions of Books as Internal Memo Warned of $10Bs Fines
Updated
Updated · TechCrunch · Jul 14
Publishers Sue Google Over Gemini Training on Millions of Books as Internal Memo Warned of $10Bs Fines
3 articles · Updated · TechCrunch · Jul 14
Summary
Hachette, Cengage, Elsevier, Scott Turow and others filed a class action in New York, accusing Google of training Gemini on copyrighted books taken from Google Books and Google Play without permission.
The complaint says Google also removed or altered copyright-management information to hide that Gemini had been trained on “stolen materials,” and cites an internal document warning the practice could trigger $10Bs-$100Bs in fines.
The case tests a narrower relationship than many AI suits because publishers had long supplied books to Google for limited search snippets, not full-scale AI training.
Two early California rulings have backed AI companies on fair-use grounds, but Anthropic was still hit with a $1.5 billion piracy penalty, leaving room for a New York judge to shape the next phase of AI copyright law.
If courts force Google to erase its AI's training, what does a 're-educated' artificial intelligence look like?
Is training AI on copyrighted books a form of learning or the largest heist in history?
Google’s Gemini AI Lawsuit: Publishers Demand Billions for Alleged Copyright Violations in AI Training Data
Overview
A major lawsuit has been filed by leading publishers and author Scott Turow against Google, accusing the company of widespread copyright infringement. The case centers on claims that Google used millions of copyrighted books and articles—often favoring high-quality works from respected authors—without permission to train its Gemini AI models. According to the report, Google allegedly bypassed proper licensing, sourcing materials from pirated sites like Z-Library and subscription services such as Scribd, all to build a vast, high-quality dataset at minimal cost. This legal battle highlights growing concerns over intellectual property rights in the age of artificial intelligence.