Anthropic’s $1.5B copyright settlement is getting messy as judge delays approval

On Thursday, a federal judge has declined to approve Anthropic’s $1.5 billion copyright settlement amid objections from authors and class members over legal fees and compensation, marking a significant delay in what was expected to be the largest copyright settlement in U.S. history.

U.S. District Judge Araceli Martinez-Olguin requested further clarification from the parties before granting final approval, citing concerns over the high attorney fees and the relatively small payouts for individual authors. Several objections argue that lawyers are requesting more than $320 million in legal fees, which critics say could equate to excessive hourly rates and unfairly reduce the compensation available to authors.

One objector, Pierce Story, estimated that lawyer fees could amount to roughly $10,000–$12,000 per hour, and argued that the settlement’s structure favors legal fees over author payouts. He suggested that reducing attorney fees to $70 million could significantly increase author compensation, which currently averages around $3,000 per author. Another objector, Ruben Lee, stated that the offered compensation is insufficient to reflect the value of their works.

Additionally, objections include concerns about the settlement’s transparency and future use of pirated works. James R. Sills, another author, emphasized the need for Anthropic to agree to destroy all copies of his works, digital or physical, before proceeding, citing ambiguity over how his works were acquired.

Why It Matters

This delay highlights ongoing tensions over the fairness of large-scale copyright settlements, particularly regarding attorney fees and fair compensation for authors. The case could set a precedent for future class-action settlements involving AI training and copyright issues, emphasizing the importance of transparency and equitable distribution of settlement funds. The objections also signal potential legal challenges that could impact the settlement’s final approval and implementation.

Background

Anthropic’s proposed $1.5 billion settlement stems from allegations that the company used a large number of copyrighted works without proper authorization to train its AI models. The settlement, announced earlier this year, was seen as a landmark in AI-related copyright disputes, with over 92% of claims filed covering more than 480,000 works. Previously, Judge William Alsup approved the settlement but recommended an independent review of attorney fees, although this was not fully disclosed to the current judge or the class members.

Objections from authors and class members have been mounting, criticizing the disproportionate legal fees and the lack of protections regarding future use of their works. A related lawsuit was filed by a group of 25 class members who opted out of the settlement, indicating ongoing resistance and legal uncertainty surrounding the case.

“Every dollar that Counsel takes from the Settlement fund is one that is not given to those actually harmed.”

— Pierce Story, author and objector

“The recommendation for an independent investigation was not fully disclosed, and class members were not notified.”

— Lea Bishop, copyright law professor and objector

“All forms of my works must be destroyed and not utilized by Anthropic until their acquisition method is clarified.”

— James R. Sills, author and objector

What Remains Unclear

It remains unclear whether the court will ultimately approve a revised settlement, what specific adjustments may be made to attorney fees, or how Anthropic will address concerns about future use of pirated works. The outcome of the ongoing objections and potential legal challenges is still uncertain.

What’s Next

The court has scheduled a response deadline for authors to address objections by May 21. Anthropic is also expected to file a brief explaining why late opt-outs should not be honored. The judge will then decide whether to approve the settlement or require modifications, potentially leading to further legal proceedings.

Key Questions

Why was the settlement delayed?

The judge delayed approval to better understand objections regarding excessive legal fees, insufficient author payouts, and concerns over future use of pirated works.

What are the main objections from authors?

Authors and class members argue that attorney fees are excessively high, payouts are too low relative to the damages, and that Anthropic should destroy all copies of their works before proceeding.

Could the settlement be rejected?

Yes, if the court finds the objections valid and the settlement terms unfair, it could reject or require modifications before final approval.

What happens if the settlement is approved?

If approved, the settlement would distribute over $1.5 billion, with a significant portion allocated to legal fees and author payouts, and set a precedent for future AI copyright disputes.