Judge approves record $1.5B Anthropic copyright settlement
The judge has signed off on the 2025 settlement.
Image by Primakov | Shutterstock
- A federal judge approved Anthropic’s $1.5 billion settlement with authors over illegally obtained books used to train Claude.
- The court said AI training on books can be fair use, but Anthropic still infringed copyrights by obtaining books illegally.
- The deal could shape future AI copyright disputes by separating lawful training use from unlawful sourcing of protected works.
Key Takeaways by nexos.ai, reviewed by Cybernews staff.
A San Francisco federal judge has greenlit a settlement requiring Anthropic to pay $1.5 billion to resolve a class-action lawsuit brought by a group of authors. The authors had accused the AI company of misusing their books to train its AI chatbot, Claude.
The case revolves around so-called shadow libraries – large databases containing copyright-protected books that are illegally distributed online.
Anthropic used these shadow libraries, which contained over 500,000 books, to train its AI model, Claude.
In August 2024, 3 authors sued Anthropic, accusing the company of using pirated versions of their works to teach Claude how to respond to human prompts.
US District Judge William Alsup ruled that Anthropic was allowed to train Claude on the works of the writers under the fair use principle, even without the writers’ permission.
However, the judge also ruled that Anthropic obtained most of the books illegally, thereby infringing the authors’ copyright.
A separate lawsuit regarding the amount of damages Anthropic had to pay for training Claude by using illegally obtained books followed.
Under US copyright law, willful copyright infringement can result in damages of up to $150,000 per work. That’s why Anthropic proposed to pay $1.5 billion to settle the lawsuit over the unlawful use of copyright-protected material.
Stay updated with our latest stories and follow us on social media
Be the first to discover new stories, ideas, and updates from our team.
Now-retired Judge Alsup initially approved the settlement in September 2025, but some of the plaintiffs objected, arguing that the proposed settlement offer wasn’t high enough.
On Monday, US District Judge Araceli Martinez-Olguin overruled their objection. She said that complaints about the settlement’s size were “not grounded in a realistic assessment of the overall risks and rewards of a trial.”
“We reached this settlement in 2025, after the court’s landmark ruling that training AI on books is fair use under copyright law, which remains the law today,” Aparna Sridhar, Anthropic’s Deputy General Counsel, said in a statement to Reuters.
According to Anthropic, more than 91% of authors and publishers covered by the settlement have claimed their share of the payment.
