Judge Rejects Anthropic’s $1.5 Billion AI Copyright Settlement

A federal judge has rejected a proposed $1.5 billion settlement between Anthropic and authors. The deal was meant to resolve a massive class-action lawsuit over copyright infringement. The ruling was issued in a California district court this week.

Judge William Alsup found the agreement fundamentally incomplete. He stated it was being forced “down the throat of authors” prematurely. The case centers on Anthropic’s use of copyrighted books to train its Claude AI model.

Key Flaws Identified in Proposed AI Payout Deal

Judge Alsup identified several critical issues with the settlement. He noted a complete lack of essential details for a fair resolution. The list of copyrighted works involved was not finalized.

Furthermore, a comprehensive list of affected authors was missing. The judge also criticized the proposed notification process for class members. According to Reuters, these omissions left the deal unworkable.

Court Mandates Strict New Timeline for Anthropic

The court has now imposed a strict new deadline for Anthropic’s legal team. Lawyers must produce a complete list of all works involved by September 15. This will allow for proper identification of all class members.

A new, clearer claim form must also be designed. It must give each author an explicit choice to opt in or out. The court must approve all final lists and the claim form by October 10 for the settlement to proceed. This ensures authors are not left behind when money is distributed.

This rejection marks a significant hurdle for Anthropic. It underscores the complex legal challenges facing AI development. The court’s demand for clarity protects the rights of authors in this pivotal AI copyright case.

Info at your fingertips

Why was the Anthropic settlement rejected?

The judge found the settlement incomplete. Key details like the list of books and authors were missing. He felt it was unfair to authors without this information.

What happens to the lawsuit now?

The lawsuit continues unless a new settlement is approved. Anthropic’s lawyers must provide the missing information by mid-September. The court will review everything again in October.

How much was each author going to get?

The original settlement was roughly $3,000 per person. This was for an estimated 500,000 authors. That payout is now on hold indefinitely.

Did the judge rule AI training is illegal?

No. The judge previously ruled training AI on books can be fair use. This case specifically involves claims of illegally downloading books from pirate sites to train the AI.

What is the main issue for AI companies?

AI companies need massive data to train their models. This often includes copyrighted material. They are now facing numerous lawsuits over how they acquire and use this data.