San Francisco, A US federal judge has granted final approval to Anthropic’s $1.5 billion settlement of a class-action copyright lawsuit brought by authors who accused the artificial-intelligence company of unlawfully obtaining copies of their books.
US District Judge Araceli Martínez-Olguín approved the settlement in the Northern District of California on July 20. The agreement covers more than 480,000 works and has been described by the plaintiffs’ lawyers as the largest copyright recovery recorded in the United States.
The lawsuit, Bartz v. Anthropic, was filed in 2024 by authors Andrea Bartz, Charles Graeber and Kirk Wallace Johnson. The plaintiffs alleged that Anthropic obtained millions of books from unauthorised online repositories while assembling a central digital library and developing its Claude AI models.
The court’s previous ruling made an important distinction between using lawfully obtained books for AI training and acquiring unauthorised copies. Then-presiding Judge William Alsup held that using books to train Anthropic’s models qualified as fair use under the circumstances examined. However, he found that storing more than seven million allegedly pirated books in a central library created a separate potential copyright liability.
Anthropic agreed to settle the claims without admitting wrongdoing. The agreement received preliminary approval in September 2025, while final approval was delayed as the court considered objections involving compensation, legal fees and the treatment of certain copyright holders.
Some authors argued that the settlement amount was inadequate or that the proposed lawyers’ fees were excessive. Judge Martínez-Olguín rejected the objections, finding that criticism of the settlement value did not realistically account for the risks and potential outcomes of taking the case to trial. The court awarded the plaintiffs’ lawyers more than $101 million in fees.
More than 91% of the eligible works were reportedly included in submitted claims. The amount ultimately received for each work will depend on the approved claims, administrative expenses and the allocation of rights between authors, publishers and other copyright holders.
The ruling does not automatically determine copyright law outside the United States or establish that every use of copyrighted material for AI training is lawful. However, it could influence licensing negotiations and data-acquisition practices across the global AI industry.
For Indian publishers, authors and AI developers, the case highlights the legal distinction between training models on legitimately obtained content and acquiring material from unauthorised sources. The US decision is not binding in India, where copyright questions involving generative AI would be assessed under Indian law.
Anthropic is a San Francisco-based AI company and the developer of the Claude family of models. Founded in 2021, the company develops general-purpose AI systems and is backed by investors including Amazon and Google parent Alphabet.
