Google sued for allegedly copying millions of books for AI training

A coalition of publishers and authors has filed a class action lawsuit against Google in a New York federal court, alleging the company unlawfully copied millions of copyrighted books and journal articles to develop and train its Gemini artificial intelligence models.

The plaintiffs, including Hachette Book Group, Cengage Learning, Elsevier, and authors Scott Turow and S.C.R.I.B.E. Inc., claim Google engaged in widespread copyright infringement. They accuse the company of copying material from its own book services, downloading works through web scraping, and repeatedly reproducing them during AI training, while allegedly removing copyright management information.

According to the lawsuit, Google allegedly utilized content provided for purposes such as discovery and search through services like Google Books and Google Scholar, repurposing it for AI training without securing new permissions. The complaint further states that Google scraped books and articles from various internet datasets, potentially including pirated sources and paywalled content, which were then allegedly copied multiple times for Gemini's training data.

Evidence cited in the lawsuit includes internal Google documents that purportedly discuss the legal risks associated with using copyrighted books for training. These documents allegedly warned of significant potential fines and acknowledged that publishers might view such use as infringement. The plaintiffs argue that Gemini now directly competes with the books used for its training, offering cheaper alternatives that could undermine sales.

The lawsuit asserts that Google's actions have harmed the market for licensed content for AI training and created AI-generated substitutes for published works. The plaintiffs contend that Gemini can generate content such as fiction for a fraction of the cost of a published book, creating an unfair competitive landscape.