Anthropic has agreed to pay $1.5 billion to settle a copyright infringement lawsuit brought by a group of authors over the use of books to train its AI models, with court approval expected to be considered next week. According to NPR, the settlement is described by the motion as the largest publicly reported copyright recovery and would compensate authors about $3,000 per book across an estimated 500,000 works.
Fair use ruling and ‘pirated’ data dispute
The case produced a split decision on fair use from U.S. Senior District Judge William Alsup in San Francisco. In a June ruling, the judge wrote that Anthropic’s training use of the plaintiffs’ books was fair use and “exceedingly transformative.” However, he determined that the company’s use of millions of books obtained from websites such as Library Genesis and Pirate Library Mirror—described as pirated copies—was not permitted, and ordered a trial on that issue before the parties announced the settlement.
NPR reports that the judge’s order stated Anthropic pirated more than 7 million copies, and noted that statutory damages for willful infringement can reach up to $150,000 per infringed work under U.S. law. Cecilia Ziniti, a tech industry lawyer not involved in the case, said the ruling effectively means that in Northern California, AI companies can train on copyrighted works if the copies are obtained legally.
Reactions from parties and creative groups
Statements underscore compensation and precedent
“Today’s settlement, if approved, will resolve the plaintiffs’ remaining legacy claims,” said Anthropic Deputy General Counsel Aparna Sridhar in a statement to NPR, adding that the company remains committed to developing safe AI systems. Justin Nelson, an attorney for the authors, called it “the first of its kind in the AI era,” saying it provides meaningful compensation and sets a precedent requiring AI companies to pay copyright owners.
Authors Guild CEO Mary Rasenberger described the agreement as a vital step toward acknowledging that AI companies cannot take authors’ work without permission, and said she expects more licensing that gives authors compensation and control. Keith Kupferschmid, president and CEO of the Copyright Alliance, said the amount represents a clear victory for publishers and authors and demonstrates that AI companies can afford to compensate rights holders.
NPR added that Anthropic this week announced a $13 billion funding round, valuing the company at $183 billion. The report also noted other ongoing legal battles, including a new lawsuit by Warner Bros. Discovery against Midjourney in California federal court, and referenced a June decision granting Meta summary judgment in a separate suit filed by Ta-Nehisi Coates and Sarah Silverman.
