A US federal judge has approved Anthropic’s $1.5 billion settlement with authors and publishers who accused the AI company of using unauthorized copies of books to train its Claude chatbot.
US District Judge Araceli Martínez-Olguin of the Northern District of California granted final approval Monday, July 20, making the agreement the largest known settlement in a US copyright case.
The settlement covers about 500,000 works, with roughly $3,000 allocated per title before legal fees and other costs. Payments will be distributed among authors, publishers, and other copyright holders with valid claims.
Anthropic said more than 91% of eligible authors and publishers covered by the agreement had filed claims. “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,” Anthropic deputy general counsel Aparna Sridhar said in a statement, according to Reuters.
Why Anthropic’s data sources became the central issue
The lawsuit began in 2024 when authors accused Anthropic of using pirated copies of their books to develop Claude, its large language model-powered chatbot. The company, backed by Amazon and Alphabet, argued that using copyrighted material to train AI models could qualify as fair use.
In June 2025, then-presiding Judge William Alsup ruled that AI training itself could fall under fair use. However, he found that Anthropic crossed a legal line by storing more than 7 million pirated books in a central library, even if those books were not necessarily used for training.
That decision left Anthropic facing potentially massive damages if the case went to trial. Under copyright law, damages could have reached up to $150,000 per willfully infringed work. The settlement avoided that trial.
A major win, but not a final answer for AI copyright
The approval marks the first major settlement in a wave of lawsuits targeting AI companies over the use of copyrighted material. Similar cases involving companies such as OpenAI, Meta, and Google remain unresolved.
Authors’ lead attorney Justin Nelson called the agreement a “historic settlement.”
“It is the largest known copyright recovery in history. We look forward to making distributions to the Class as promptly as possible,” Nelson said in a statement.
However, the settlement does not create a nationwide legal standard for AI training. Because Anthropic settled before an appeal, the earlier fair-use ruling remains a district court decision rather than a binding precedent.
What it means for the AI industry
The settlement gives AI companies a clearer warning: how they obtain training data matters as much as how they use it. The case suggests that courts may be more open to allowing AI systems to learn from copyrighted works. However, companies could still face serious consequences if their data collection methods involve unauthorized sources.
For AI developers, the next challenge will be building training datasets that are large enough to compete while avoiding legal uncertainty. For creators, the settlement provides compensation but does not fully resolve concerns about how their work is used in future AI systems.
Some authors and publishers rejected the agreement and have chosen to continue pursuing separate lawsuits against Anthropic. Those cases could still shape the long-term rules governing AI and copyright.
Related news: Major publishers bring another AI copyright lawsuit against Google as copyright holders continue to challenge how AI companies source and use training data.