Anthropic’s Landmark $1.5 Billion Copyright Settlement Receives Final Approval

Anthropic can now begin compensating authors and publishers after a federal judge granted final approval to the AI company’s landmark $1.5 billion copyright settlement. The class action lawsuit accused the AI lab of infringing copyright by downloading and storing millions of protected books to build its training library.

The settlement was originally given preliminary approval by Judge William Alsup of the U.S. District Court for the Northern District of California after he ruled that Anthropic had unlawfully downloaded and retained copyrighted books. Following Alsup’s retirement, Judge Araceli Martinez-Olguin issued the final approval, clearing the way for payments to rights holders.

Authors to receive payments for around 500,000 works.

The agreement provides roughly $3,000 for each of an estimated 500,000 works, with payments shared among the authors and publishers that own the copyrights. Although the settlement is widely regarded as the largest in U.S. copyright history, many creators remain dissatisfied because the court ruled that using copyrighted books to train AI models qualifies as fair use.

However, the judge distinguished between using copyrighted material for AI training and how Anthropic obtained part of its library. While books purchased and scanned by the company were considered lawful, downloading books from pirate websites such as Library Genesis and Pirate Library Mirror was ruled illegal. Anthropic agreed to settle before the piracy claims reached trial, avoiding the risk of potentially larger damages.

Legal debate over AI training continues.

The final approval closes this case but does not establish a binding legal precedent because the dispute ended before it could reach an appeals court. Other courts remain free to reach different conclusions on similar copyright claims involving AI developers.

Several major technology companies, including Google, Meta, Midjourney and OpenAI, continue to face lawsuits over whether copyrighted material can legally be used to train AI models. Most recently, a group of publishers and authors, including Hachette, Cengage, Elsevier, Scott Turow and S.C.R.I.B.E., filed a class action lawsuit accusing Google of using copyrighted books to train its Gemini AI platform without permission.TechCrunch first published exclusive reporting on the original story.