Anthropic to Pay $1.5 Billion in Landmark Copyright Settlement with Authors and Publishers

San Francisco – Artificial intelligence company Anthropic has agreed to pay $1.5 billion to settle a class-action lawsuit brought by book authors and publishers, marking the largest copyright recovery in history within the AI era. The agreement, which awaits approval from U.S. District Judge William Alsup, resolves claims that Anthropic illegally acquired and used millions of pirated books to train its Claude chatbot. This significant settlement could set a precedent for ongoing legal battles between AI developers and creative professionals.

The lawsuit, initiated by authors Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson, alleged that Anthropic's training practices constituted "large-scale theft" of copyrighted material. While a June ruling by Judge Alsup found that training AI models on copyrighted works could be considered fair use if legally obtained, he also determined that Anthropic had wrongfully downloaded over 7 million digitized books from pirate websites like Library Genesis and Pirate Library Mirror. This distinction between legitimate acquisition and pirated sources was central to the case.

Under the terms of the settlement, Anthropic will pay approximately $3,000 for each of an estimated 500,000 books covered by the agreement. Justin Nelson, a lawyer for the authors, stated that this is "the first of its kind in the AI era," emphasizing the historical significance of the payout. Anthropic has also committed to destroying the original pirated book files it downloaded, addressing a key concern of the plaintiffs.

Aparna Sridhar, Anthropic's deputy general counsel, affirmed that the settlement, if approved, "will resolve the plaintiffs' remaining legacy claims." She reiterated the company's commitment to developing safe AI systems. Mary Rasenberger, CEO of the Authors Guild, hailed the outcome as "an excellent result for authors, publishers, and rightsholders generally," sending a strong message to the AI industry regarding the consequences of pirating creative works.

Experts suggest that avoiding a trial, which was scheduled for December, prevented potentially higher damages that could have crippled the San Francisco-based company. This settlement is expected to influence other high-profile copyright lawsuits against AI firms, including those involving OpenAI and Meta, as the industry grapples with establishing legal frameworks for data acquisition and AI model training.