Artificial intelligence startup Anthropic has agreed to pay at least $1.5 billion to settle a U.S. class action lawsuit over claims it used pirated books to train its AI models, marking one of the largest copyright settlements in history.
The lawsuit, brought by authors Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson, accused Anthropic of copying their works without permission to develop Claude, its AI chatbot that competes with OpenAI’s ChatGPT. According to court documents, the settlement covers roughly 500,000 books, translating to around $3,000 per title — four times the minimum statutory damages under U.S. copyright law.
“This landmark settlement far surpasses any other known copyright recovery,” said plaintiffs’ attorney Justin Nelson. “It is the first of its kind in the AI era.”
While Anthropic secured a partial legal victory earlier this year, when U.S. District Court Judge William Alsup ruled that training AI models on books constituted “fair use” under copyright law, the court also found the company’s practice of downloading millions of pirated works to create a permanent digital library was not protected. “The technology at issue was among the most transformative many of us will see in our lifetimes,” Judge Alsup wrote, but added that blanket protection could not extend to wholesale piracy.
Under the settlement, Anthropic will destroy pirated files and copies while retaining rights to books it legally purchased or scanned. “We remain committed to developing safe AI systems that help people and organisations extend their capabilities, advance scientific discovery, and solve complex problems,” said Aparna Sridhar, Anthropic’s deputy general counsel.
The deal comes as Anthropic cements its place among the world’s leading AI firms. The San Francisco-based company recently raised $13 billion in funding, valuing it at $183 billion, as it competes against OpenAI, Google, Meta and Microsoft in the rapidly accelerating AI race.
Mary Rasenberger, CEO of the Authors Guild, said the agreement sends a clear warning to the industry. “This settlement sends a strong message to the AI industry that there are serious consequences when they pirate authors’ works to train their AI, robbing those least able to afford it,” she said.
The case adds to a growing list of legal battles confronting AI companies. In June, Meta secured a victory when a San Francisco judge ruled its use of copyrighted material to train its Llama AI system qualified as “fair use.” Apple, meanwhile, is facing its own lawsuit from authors Grady Hendrix and Jennifer Roberson, who allege the company used pirated works to train “Apple Intelligence,” its suite of new generative AI tools.
The Anthropic settlement still requires court approval but is expected to set a precedent for how copyright law will be applied to AI training in the years ahead.
