Anthropic’s $1.5 billion settlement for an AI copyright lawsuit has been approved by a federal judge in San Francisco. The lawsuit was filed by a group of writers in 2024, accusing Anthropic of using pirated copies of their work to train its Claude large language models.
William Alsup, the original judge of the case, ruled that while Anthropic’s training of Claude constituted fair use, the company infringed copyright by storing 7 million pirated books in a central library. He rejected a proposed settlement in September 2022, citing concerns that the terms were inadequate and not addressing significant questions.
Following the rejection, Anthropic established an official settlement website, listing affected works for authors to verify their eligibility for claims. The approved settlement now encompasses over 480,000 works.
US District Judge Araceli Martinez-Olguin, succeeding Alsup, has approved the settlement terms. Authors are anticipated to receive around $3,000 for each affected work, while Anthropic is mandated to destroy the pirated copies in its possession. Some plaintiffs have opted out of the settlement, deeming the payout insufficient, and plan to pursue separate lawsuits against the company.
