Two Stanford authors rejected a $1.5 billion settlement and are suing Anthropic over pirated books used to train Claude, a fight that could shape AI costs.
Two Stanford authors are suing Anthropic, the company behind the AI chatbot Claude. They say it trained Claude on pirated copies of their books. They turned down a $1.5 billion settlement in a larger case and want a jury to decide instead. They are seeking up to $150,000 per book, far above the roughly $3,000 per work the settlement offered, according to The Stanford Daily.
In the earlier case, Bartz v. Anthropic, Judge William Alsup drew a line. Anthropic could not download pirated books to build a permanent central library. But he treated training Claude on books obtained legally as fair use. Fair use is a legal idea that allows limited use of copyrighted work without permission. He compared AI training to a person who reads a book and absorbs its themes and style.
The settlement went to authors and publishers of nearly 500,000 registered works. The Authors Alliance calls it the largest copyright settlement involving an AI company. Tobias Wolff, a Stanford professor emeritus of English and National Medal of Arts winner, and English professor Chang-rae Lee chose to opt out. Wolff says 10 of his publications were used without permission. He says the settlement still gives Anthropic "pretty free rein" over literature.
Their new suit is called Cruz v. Anthropic and was filed in May 2026. It claims Anthropic downloaded books from "shadow libraries" such as LibGen and PiLiMi, which are pirate collections, and kept the files in a central library. If the authors win, they would gain a permanent order barring Claude training on their works.
This is one of many fights. Authors, visual artists and news outlets, including The New York Times and Getty Images, have filed over 70 infringement lawsuits against AI companies. The Authors Guild is suing OpenAI and Microsoft and is working with Congress on national legislation that would require AI companies to license the works they use.
Stanford Law scholar Paul Goldstein notes the practical problem. Training can copy millions of works, and getting a license for each one is, in his words, impracticable.
The material does not say when Cruz v. Anthropic might go to trial or how Anthropic has responded. The legal future of mass scraping for AI training remains uncertain. For businesses, the outcome may decide whether AI tools must pay for the books and articles they learn from. That could affect the prices and rules attached to those tools.
Courts are still deciding whether AI makers must pay for the books they learn from, and the answer could change what AI tools cost and allow.
Lawmakers in both parties want Nvidia's CEO to answer questions about a reported tie to an alleged Chinese intelligence asset, and chip rules could shift as a result.
Read more: Lawmakers from both parties want Nvidia's CEO to answer questionsDeepSeek says 1.54 million people use its AI agent platform each day. If you run a business, cheaper AI helpers may soon be offered to you.
Read more: DeepSeek says its AI helper platform now has 1.54 million daily usersOpenAI has about $11 billion in hand at a $1.25 trillion value, below its first ask. Microsoft and NVIDIA shareholders have reason to watch what comes next.
Read more: OpenAI raises $11 billion, but at a lower price than it wanted
