Anthropic reaches historic $1.5 billion settlement with authors in landmark AI copyright lawsuit

Anthropic has reached a $1.5 billion settlement with a group of authors who accused the artificial intelligence company of using their books without authorization to train its Claude chatbot. The class-action settlement was approved by a federal court in San Francisco.

The agreement marks the first major lawsuit among dozens filed by rights holders against technology companies in the US to resolve with a significant settlement.

The authors initiated the lawsuit in 2024, accusing Anthropic of using pirated versions of their books to train its AI models without securing the necessary permissions.

According to a report by Reuters, Aparna Sridhar, Anthropic’s Head of Issues and Policy, said in a statement: “We reached this agreement in 2025. The settlement follows a landmark court ruling, which remains valid today, establishing that training artificial intelligence on books constitutes fair use under copyright law.”

Justin Nelson, an attorney representing the plaintiff authors, described the development as a “historic settlement.” Nelson added that the agreement reached with Anthropic could be considered the largest monetary payout in the history of copyright law.

Meanwhile, some authors and publishers declined to participate in the class-action lawsuit, choosing instead to file independent lawsuits against Anthropic. The judicial processes for these individual cases against the company are ongoing.

Prior to this development, Anthropic filed a lawsuit against the Pentagon in March to challenge an attempt by the US Department of Defense to blacklist the company on national security grounds.

In June, the US government decided to block foreign users from accessing the company’s most advanced AI models, Fable 5 and Mythos 5.

David Sacks, a US investor and Co-Chair of the President’s Council of Advisors on Science and Technology, explained that the restriction was implemented after it was discovered that the integrated safety mechanisms within the models could be bypassed.

Two weeks after that restriction was imposed, the US government restored access to the most powerful model, Mythos 5, for select American entities, including major corporations and government agencies.