A U.S. federal court has put an end to one of the most high-profile intellectual property cases in the field of artificial intelligence. Anthropic, the developer of the famous chatbot Claude, has officially finalized a settlement with authors totaling $1.5 billion. This is the largest payout in the history of U.S. copyright litigation, and it sets a new standard for the entire industry.

How the Conflict Developed

The lawsuit was filed back in 2024. A group of authors accused Anthropic of mass downloading pirated copies of books to train its language model. Judge William Alsup initially recognized the process of training AI on texts as fair use. However, he considered the storage of over 7 million unlicensed files a direct violation of the law. After Alsup's retirement, the case was transferred to Judge Araceli Martinez-Olguin, who issued the final verdict, approving the terms of the settlement.

"The agreement was reached in 2025 after the court recognized that training AI on books falls under fair use under copyright law—and this standard still applies," said Aparna Shridhar, Deputy General Counsel at Anthropic.

Figures and Terms of the Deal

According to court documents, authors will receive approximately $3,000 per work. The list includes more than 480,000 titles. Judge Martinez-Olguin specifically emphasized that this amount is four times the minimum payout of $750 for a standard copyright violation. By April, rights to payments were confirmed for 440,490 works—91.3% of the entire list. Additionally, Anthropic is required to destroy all illegal copies of the books.

The judge rejected objections to the $1.5 billion compensation amount, calling it fair and proportionate. However, some authors declined to participate in the collective agreement and continue to litigate against the company separately.

My analysis: This is the first major settled case of copyright infringement by AI in the U.S., and it creates a powerful precedent. For the crypto industry, where AI technologies are increasingly integrated into smart contracts, DeFi protocols, and analytics platforms, this case is a warning signal. If you use data to train models, be prepared for legal risks and financial reserves. $1.5 billion is not a fine—it is the price of neglecting copyright in the era of generative AI.