Anthropic $1.5 billion AI settlement sparks rights disputes

Authors and publishers are engaged in disputes over the distribution of a $1.5 billion settlement with Anthropic, the technology company behind the AI chatbot Claude. The settlement follows a class-action lawsuit in the U.S. District Court for the Northern District of California, which found that Anthropic used pirated books to train its models.

The agreement covers more than 482,000 works, with potential payments of up to $3,000 per illegally downloaded book. However, a settlement administrator is facing conflicting claims from authors and publishers regarding who holds the rights to specific titles. Mary Rasenberger, CEO of the Authors Guild, noted that these discrepancies often arise because some publishers may not maintain accurate records of books whose rights have reverted to authors.

Under the terms of the settlement, proceeds must be shared between authors, co-authors, and publishers. For textbooks, the author's share is typically lower, ranging from 10% to 15% depending on individual contracts. While some authors have successfully proven ownership to receive full payouts, the complexity of rights management remains a central challenge in the distribution process.