Anthropic Ordered to Pay $1.5 Billion in Record-Breaking AI Copyright Case

A Landmark Ruling for Authors and the AI Industry

A U.S. federal judge has approved a $1.5 billion settlement between Anthropic and a group of authors, closing the largest copyright class action in U.S. history and marking the first major resolution of a wave of lawsuits against AI companies over the use of copyrighted books in model training. The ruling, issued on July 20, sends a thunderous signal to Silicon Valley: using pirated content to build artificial intelligence may come with a staggering price tag.

U.S. District Judge Araceli Martínez-Olguín approved the deal on July 20 in San Francisco, rejecting objections from some authors who argued the payout was too small. The settlement resolves a class action filed by writers Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson in August 2024, who accused Anthropic of using pirated books to train its Claude chatbot. What started as a complaint from three authors quickly ballooned into one of the most consequential legal battles in the history of intellectual property.

How Anthropic Built Claude — and Who Paid the Price

Believed to be the largest publicly reported recovery in the history of U.S. copyright litigation, the massive payout sends a powerful message of accountability to AI developers who torrented copyrighted works from illegal pirated websites to train AI models. Anthropic's acquisition and use of books from these websites — illegal websites that have been repeatedly shut down by law enforcement and the courts — was made public for the first time through this litigation.

The class certification turned three individual authors into representatives of thousands of rightsholders in a mega-lawsuit representing nearly half a million works. With statutory damages potentially reaching $150,000 per work, Anthropic was suddenly staring down the barrel of theoretical liability exceeding $70 billion. Settling for $1.5 billion, while still an enormous sum, was arguably the company's most rational move.

Under the deal, authors and publishers receive $3,000 for each of an estimated 500,000 works covered by the settlement. More than 91% of those eligible have already filed claims, according to Anthropic. Beyond the money, Anthropic will also be required to destroy the original files of the works downloaded from pirated book datasets within 30 days of the final judgment and certify that the Library Genesis and Pirate Library Mirror datasets with pirated material were used in training any of the company's commercially released large language models.

What the Settlement Does — and Doesn't — Settle

The most consequential thing the ruling settled may be what it did not settle. The fair-use ruling that preceded it — issued by then-presiding Judge William Alsup — was a single district-court decision. Because Anthropic chose to settle rather than let the case proceed to an appeals court, that ruling will never become binding precedent. Every other AI company facing similar lawsuits remains in legal limbo, with no higher-court guidance on whether training AI on copyrighted material constitutes fair use.

The settlement addresses Anthropic's past infringements, does not give Anthropic permission for future use of copyrighted works, and emphasizes the need for AI companies to move toward a licensed, permission-based access business model. That last point may prove to be the ruling's most lasting legacy — not as legal precedent, but as market pressure.

A New Era for AI and Creative Rights

"This settlement marks the beginning of a necessary evolution toward a legitimate, market-based licensing scheme for training data," said Cecilia Ziniti, a tech industry lawyer and former Ninth Circuit clerk who followed the case closely. The creative community has watched the AI boom with growing alarm, and this settlement — however imperfect — represents the first concrete proof that the courts can deliver accountability.

The proposed deal marks the first settlement in a string of lawsuits against tech companies including OpenAI, Microsoft, and Meta Platforms over their use of copyrighted material to train generative AI systems. With those cases still pending, the Anthropic ruling gives plaintiffs' attorneys a powerful data point and gives AI companies a very expensive reason to reconsider how they source their training data going forward.