A federal judge approved a $1.5 billion settlement requiring AI company Anthropic to compensate authors whose books were used without permission to train its Claude chatbot.

More than 482,000 books are covered by the agreement, with authors and publishers receiving about $3,000 per claimed work.

The settlement follows a ruling that found AI training itself may qualify as fair use, but Anthropic improperly obtained millions of books from pirated sources.

Anthropic said it welcomed the resolution while maintaining that AI training on books is protected under copyright law, the AP has reported.

The case marks the first major settlement among numerous ongoing copyright lawsuits involving artificial intelligence companies.