Publishers sue Google over alleged unauthorized use of copyrighted works to train Gemini

A coalition of publishers and authors has filed a class action lawsuit against Google, accusing the company of training its Gemini AI models on copyrighted works without permission. The plaintiffs include Hachette, Cengage, Elsevier, author Scott Turow, and the Society for Civil Reading and Independent Book Exchanges (S.C.R.I.B.E.). The lawsuit alleges that Google intentionally removed or altered copyright information to conceal that its models were trained on 'stolen materials.'

The complaint centers on Google’s use of copyrighted books obtained through programs such as Google Books and Google Play, which the publishers argue were designed for limited purposes like searchable snippets and bibliographic access. According to the lawsuit, Google copied these works for AI training despite lacking authorization. The plaintiffs claim this use was neither agreed upon nor compensated.

The lawsuit also cites an internal Google document that reportedly warned that using copyrighted books for AI training could be 'highly problematic' and might result in potential fines ranging from $10 billion to $100 billion. Google has not publicly responded to requests for comment.

This case is one of many lawsuits filed by publishers, authors, and copyright holders against AI companies, including Google, Meta, OpenAI, and Anthropic. While two early court decisions in California ruled that AI training on copyrighted works constitutes 'fair use,' this lawsuit was filed in the U.S. District Court for the Southern District of New York, offering a new venue and judge the opportunity to weigh in. Legal experts note that these rulings do not establish binding precedent, leaving the issue unresolved.

The outcome of this lawsuit could influence how AI companies approach the use of copyrighted material in training data, potentially reshaping industry practices and legal strategies around fair use and licensing.