A federal judge has given final approval to Anthropic's $1.5 billion settlement with a group of authors, marking the largest known payout in a US copyright case. US District Judge Araceli Martinez-Olguin signed off on the deal Monday, overruling objections that the sum wasn't big enough. The settlement resolves claims that the AI company used pirated copies of books to train its Claude chatbot.
The largest copyright payout in US history
Authors sued Anthropic in 2024, alleging the company trained Claude on bootleg versions of their work. The case was one of the first to test how copyright law applies to generative AI. Retired Judge William Alsup ruled last June that training AI on books could be fair use, but storing more than 7 million pirated copies was illegal. That ruling set the stage for the settlement.
Under the deal, authors will receive about $3,000 per work for more than 480,000 titles. That's four times the $750 minimum for standard copyright infringement. As of April, claimants had covered 440,490 works, or 91.3% of the list. The settlement also requires Anthropic to destroy the pirated book files it used for training.
Why the judge rejected objections
Several objectors argued the $1.5 billion figure was too low given the scale of the infringement. Judge Martinez-Olguin disagreed. She called the objections 'not grounded in a realistic assessment of the overall risks and rewards of a trial.' The judge noted that the settlement avoided a lengthy, uncertain trial and delivered compensation far above the statutory minimum.
The deal is the first major US AI copyright dispute to settle. It could set a template for other cases, though the facts here were specific: a clear admission of storing pirated files, a ruling that distinguished between training and storage, and a company with deep pockets.
What the settlement requires
Anthropic must destroy the pirated book files within a set timeframe. The company also agreed to pay the $1.5 billion into a fund for claimants. Authors who didn't opt out will receive payments based on the number of their works used. The settlement doesn't impose any restrictions on how Anthropic can train future models, as long as it uses lawfully obtained material.
Not all authors are satisfied. Some opted out of the settlement and continue separate lawsuits against Anthropic. Those cases will now proceed, likely testing the same fair-use arguments that Judge Alsup addressed. The company faces ongoing scrutiny over its data sourcing practices, and the unresolved lawsuits mean the legal fight over AI and copyright isn't over yet.
