A federal judge in California has approved the largest copyright settlement in American history, ordering AI company Anthropic to pay $1.5 billion to authors and publishers whose books were used without permission to train its chatbot systems.
U.S. District Judge Araceli Martínez-Olguín ruled Monday that the massive payout was “fair, reasonable and adequate” compensation for roughly 54 known copyright holders and more than 500,000 potential class members. Each creator will receive approximately $3,100 per copyrighted work—covering about 482,000 books total.
How Anthropic Built Its AI on Pirated Books
The lawsuit accused Anthropic of knowingly training its large language models on datasets downloaded from websites notorious for pirated content, including Library Genesis and Pirate Library Mirror. These sites host hundreds of thousands of copyrighted books uploaded without author or publisher consent.
Plaintiffs argued the San Francisco-based AI company deliberately chose these sources to build its commercial products while avoiding licensing fees that would have cost far more than the settlement. The class includes anyone holding copyright to books that appeared in the specific versions of these piracy sites that Anthropic downloaded.
Victory Wasn’t Guaranteed, Judge Says
In her approval order, Judge Martínez-Olguín acknowledged the legal uncertainty facing the authors. “Success at trial was not assured, and a loss would have left the class with no recourse,” wrote the Obama appointee.
The comment reflects the unsettled state of copyright law as it applies to AI training. Tech companies have argued that ingesting copyrighted material to train algorithms constitutes “fair use” under existing law. Authors counter that AI systems built on their stolen work compete directly with them in the marketplace.
The settlement includes over $100 million in attorney fees. That leaves roughly $1.4 billion for the writers, journalists, and publishers whose livelihoods depend on copyright protection—protections that were supposed to prevent exactly this kind of mass appropriation.
The case signals potential exposure for other AI companies. OpenAI, Google, and Meta all face similar lawsuits from authors and news organizations alleging their models were trained on pirated or improperly licensed content.
Key Points
- Anthropic will pay $1.5 billion to authors whose books were used without permission to train AI models
- Each copyright holder receives roughly $3,100 per work from a class covering 482,000 pirated books
- The settlement is the largest copyright payout in U.S. history and could expose other AI companies to similar claims
https://www.courthousenews.com/anthropic-to-pay-1-5-billion-copyright-settlement-to-authors-publishers/ – July 21, 2026
