A group of authors sued Anthropic, the company behind Claude (an artificial intelligence model), arguing that it had taken their books without permission and used them to develop artificial intelligence. Anthropic had downloaded millions of books from pirated websites and created a large internal library of them.
The authors sued Anthropic in 2024, arguing that the company, backed by Amazon (AMZN.O) and Alphabet (GOOGL.O), had used pirated versions of their books without permission to teach Claude how to respond to people's requests.
The case became particularly important because the court later made a distinction between the use of books for training an AI system and how these books had been obtained illegally.
The judge ruled that training the AI with the books could be considered fair use, but Anthropic's obtaining and storing pirated copies of the books constituted copyright infringement.
Rather than take the case to a full trial, Anthropic agreed to a $1.5 billion settlement, which was finally approved in July 2026. The agreement covers about 482,000 books, with eligible rights holders expected to receive about $3,000 per work.
About 91% of the works included in the settlement have already been claimed by authors or publishers. The settlement is now considered the largest known copyright recovery in U.S. history.
However, this agreement does not resolve the much broader issue of whether AI companies should be allowed to use copyrighted books, articles, artwork, and other creative materials to train their models.
In this particular case, the court actually considered training models with legally obtained books to be fair use. The problem was that Anthropic had obtained the materials through piracy.
The settlement also drew opposition from some authors, who argued that its amount was not large enough, that it overcompensated plaintiffs' lawyers, or that it unfairly excluded some copyright holders.
Judge Martinez-Olguin rejected these objections in her ruling and said that the complaints about the settlement amount “were not based on a realistic assessment of the overall risks and benefits of a trial,” and awarded the lawyers more than $101 million of the $187.5 million they had requested in fees.
Several authors and publishers chose not to be included in the agreement and have filed separate lawsuits against Anthropic, which are still pending.
Therefore, the legal battle between AI creators and companies is still far from over.
