Key facts
- Anthropic agreed to a $1.5 billion settlement to resolve a class-action lawsuit by authors.
- Authors will receive approximately $3,000 per work for an estimated 500,000 books used to train Anthropic's AI.
- The University of Tennessee has filed a separate lawsuit against Anthropic for patent infringement.
- The patent infringement lawsuit concerns neural network technology.
- The settlement aims to resolve claims that Anthropic used pirated books to train its AI chatbot, Claude.
Artificial intelligence company Anthropic has agreed to pay $1.5 billion to settle a class-action lawsuit brought by authors who claim the company used pirated copies of their works to train its AI chatbot, Claude. The settlement, which received final approval from a judge, could set a precedent for legal battles between AI firms and creative professionals.
Anthropic will pay authors or publishers approximately $3,000 for each of an estimated 500,000 books included in the settlement. This resolution comes as the University of Tennessee has filed a separate lawsuit against Anthropic, alleging patent infringement related to neural network technology.
Authors Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson led the copyright lawsuit, arguing that Anthropic downloaded millions of books from pirate websites like Library Genesis and the Pirate Library Mirror to build its AI models. A previous ruling by U.S. District Judge William Alsup found that while training AI on copyrighted books might not be illegal, Anthropic had wrongfully acquired millions of them.
Anthropic stated that the settlement resolves the plaintiffs' remaining legacy claims and reiterated its commitment to developing safe AI systems. As part of the agreement, the company will also destroy the original book files it downloaded. Legal analysts suggest that without the settlement, Anthropic could have faced significantly higher costs if the case had gone to trial.