Anthropic's $1.5bn copyright settlement gets final court approval

What happened: US District Judge Araceli Martínez-Olguín gave final approval on Monday 20 July 2026 to Anthropic's $1.5 billion settlement with authors and publishers over books used to train Claude.

The numbers: Roughly $3,000 per work, spread across about 500,000 works. Payments to authors begin now. It's believed to be the largest copyright settlement in US history.

The context: Anthropic sourced its training books two ways: buying and scanning physical copies, and downloading titles from pirate sites. Courts had already ruled that training AI on copyrighted text counts as fair use. But obtaining books via piracy did not. Rather than fight that piracy claim at trial, Anthropic settled.

Why it matters: The case puts a concrete price tag on pirated training data. It signals to AI labs that licensed or purchased content, not pirated downloads, is the safer path forward.

The catch: The settlement resolves this case only. It sets no binding precedent for the wider AI industry.

The rivals: Copyright suits over AI training data remain pending against Google, Meta, Midjourney and OpenAI. In July, publisher Hachette and author Scott Turow separately sued Google over data used to train Gemini.

What's next: Authors covered by the roughly 500,000 works begin receiving payments, while the unresolved cases against other major AI labs continue through the courts.