A federal court in San Francisco has finally approved a $1.5 billion settlement agreement between Anthropic and a class-action lawsuit filed by book authors. The plaintiffs accused the developer of the Claude chatbot of illegally using their works to train its language model. This is the largest known copyright settlement in U.S. history.

Judge Araceli Martínez-Olguín approved the deal, rejecting objections that the payments were insufficient. The lawsuit was filed in 2024: authors alleged that Anthropic used pirated copies of their books without permission to train Claude. Earlier, another judge, William Alsup, had preliminarily approved the agreement but issued a contradictory ruling: training AI on copyrighted books was deemed fair use, but storing more than seven million pirated books in a "central library" was considered a violation.

Details of the Agreement and Reactions

Anthropic's Deputy General Counsel Aparna Shridhar confirmed that the deal was reached in 2025 following Alsup's ruling. According to her, more than 91% of eligible authors and publishers have already requested their share of the payments. Lead plaintiffs' attorney Justin Nelson called the agreement historic. However, some authors objected, disputing the amount, the size of attorney fees, and the exclusion of certain rights holders. The judge rejected these claims, awarding attorneys more than $101 million of the requested $187.5 million.

Not all authors and publishers agreed to the terms: some opted out of the deal and filed separate lawsuits against Anthropic. Those proceedings are ongoing.

My analysis: This precedent sets a new standard for the entire AI industry. The $1.5 billion sum is not just compensation but a signal to the market: using pirated content to train models carries enormous financial risks. However, the court's ruling on fair use for the training itself leaves a loophole that could spark new disputes. Companies will either have to negotiate with rights holders or fundamentally change their approach to data collection.