The AI research lab Anthropic has agreed to pay $1.5 billion to a group of authors and publishers after a judge ruled that the company illegally downloaded and stored millions of copyrighted books.
The proposed settlement comes out to $3,000 per work for around 500,000 authors — the largest payout in the history of U.S. copyright cases.
The agreement comes after a U.S. District Court judge sided with Anthropic in June, ruling that when the company acquired copyrighted books legally, the law allowed it to train AI models using the material because the process transformed them into something new.
“Like any reader aspiring to be a writer, Anthropic’s LLMs trained upon works not to race ahead and replicate or supplant them — but to turn a hard corner and create something different,” Judge William Alsup wrote.
But he found that Anthropic, which makes the Claude chatbot, illegally acquired millions of books through “shadow libraries,” such as Library Genesis and Pirate Library Mirror, knowing that they contained pirated material.
If the settlement is approved by the court, Anthropic will destroy the material it illegally downloaded.
“This result is nothing short of remarkable,” the settlement reads, and “will set a precedent of AI companies paying for their use of pirated websites.”
Anthropic pointed to the June ruling that found that its approach to training AI models constitutes fair use.
“Today’s settlement, if approved, will resolve the plaintiffs’ remaining legacy claims,” said Aparna Sridhar, Anthropic’s deputy general counsel. “We remain committed to developing safe AI systems that help people and organizations extend their capabilities, advance scientific discovery, and solve complex problems.”