Anthropic's $1.5B Copyright Settlement

Anthropic has agreed to a $1.5 billion settlement in a class-action lawsuit with authors who claimed the AI company used pirated books to train its Claude chatbot. The settlement could be a turning point for legal battles between AI companies and copyright holders. The agreement resolves claims that Anthropic illegally used copyrighted material, including from pirate sites like Library Genesis and Pirate Library Mirror.

Under the settlement, Anthropic will pay approximately $3,000 per work, covering an estimated 500,000 titles. The company has also committed to destroying pirated copies of books used in the training process. The lawsuit alleged Anthropic downloaded over seven million digitized books, knowing they were pirated, to train its large language models.

The settlement is considered a landmark victory for authors and copyright holders, potentially setting a precedent for AI companies to respect copyright laws when training AI models. While a judge had ruled that training AI on copyrighted books was not illegal, the acquisition of books through pirate websites was deemed unlawful. The settlement awaits court approval.