A group of publishers and authors have filed a class-action lawsuit against Google, accusing it of using copyrighted works without permission to train the AI platform Gemini. Plaintiffs include Hachette Livre, Cengage Group, Elsevier, author Scott Turow, and S.C.R.I.B.E., among others. The lawsuit also accuses Google of intentionally deleting or modifying copyright information in works to "conceal the fact that the Gemini model was trained on stolen materials."

Google is accused of using books from Google Books and the Play Store to train AI

The lawsuit documents state that publishers and authors had long provided copyrighted works to Google with the intention of allowing books to be searchable through Google Books. However, the search results only provided limited text excerpts and table of contents, preventing users from viewing entire books. Nevertheless, the plaintiffs believe that Google later used these book copies to train the Gemini model, and also used content from books uploaded to the Google Play Store for training, despite never obtaining proper authorization. Internal Google documents cited by the plaintiffs show that Google itself recognized that using copyrighted books to train AI could lead to "serious issues" and face "potential fines ranging from hundreds of millions to billions of dollars."

Defense of "fair use" remains controversial, Anthropic fined $1.5 billion

This lawsuit is one of many initiated by copyright holders against AI companies, with Google, Meta, OpenAI, and Anthropic all facing similar disputes. Earlier rulings in California courts favored AI companies, ruling that using copyrighted content to train AI models could be considered "fair use." However, Anthropic was previously fined $1.5 billion for illegally using copyrighted works to train AI, becoming the largest copyright settlement in U.S. history, with about 50,000 authors eligible to receive at least $3,000 each. This lawsuit against Google was filed in the U.S. Southern District Court of New York, meaning another judge will have the opportunity to express a different opinion on whether AI training constitutes fair use. Google has not yet responded to requests for comment.