Over 100 Authors Sue Anthropic, Seek $75 Million Over Alleged Use of Pirated Books

Over 100 authors have sued Anthropic, seeking $75 million over allegations that the AI company used pirated books to train its models without permission.
on Jul 07, 2026

Over 100 writers have now sued the artificial intelligence firm Anthropic for using stolen books to train their artificial intelligence models without authorization and payment. This lawsuit follows the ongoing debates about copyrights, artificial intelligence, and creative materials used in artificial intelligence training that have been ongoing lately.

This lawsuit was filed in the Northern California District Court on June 17. According to the lawsuit, Anthropic stole millions of unlicensed digital copies of books from piracy sites like Pirate Library Mirror and Library Genesis. This information was apparently obtained via BitTorrent and included in Anthropic’s central library of training materials.

According to the authors, these books were then used to develop and train Anthropic's artificial intelligence system and language models for its chatbot Claude. This lawsuit states that the company used copyrighted materials without the awareness, permission, or payment to the writers who authored those works.

More than 500 books are estimated to have been involved in this lawsuit. Some of the works mentioned include the book "Get Good with Money" written by Tiffany Aliche and "Like Water for Chocolate" by the Mexican writer, Laura Esquivel. Writers who filed this lawsuit also include screenwriter Zachary Sklar, video game founder Nolan Bushnell, and Donna Barba Higuera who is an award-winning writer.

They are demanding statutory damages of $150,000 per work that was supposedly pirated. According to the lawyers representing the plaintiff, this is not an issue of innovation but whether AI firms can manufacture products based on the creative works without getting permission from the owners of those works.

This legal proceeding is in addition to another huge copyright case that was settled previously against Anthropic. Last September, the firm agreed to pay $1.5 billion to settle a legal dispute that was filed against it by authors claiming that millions of pirated digital books were used to train its language models.

The new lawsuit clearly demonstrates that issues related to the training data for AI are far from being resolved. As AI technologies become more commonplace in publishing, media, education, and entertainment industries, there is a growing call for establishing clear guidelines concerning the consent and licensing.

The decision in this case may have implications not only for the publishing industry but also for all companies utilizing AI based on text datasets.