Anthropic's proposed $1.5 billion settlement with authors is not moving forward yet. At a hearing Monday, US District Judge William Alsup sharply questioned whether the deal was ready for approval, saying it risked being pushed “down the throat of authors” before basic details were settled.
The case centers on books illegally downloaded by Anthropic to train AI. The proposed agreement could resolve a major copyright class action, but Alsup said the current version leaves too many important questions unanswered.
Why the judge refused to approve the deal
Alsup denied preliminary approval because, in his view, the settlement was incomplete. In an order, he said he was “disappointed that counsel have left important questions to be answered in the future.”
The unresolved items were not small administrative details. The Works List, the Class List, the Claim Form, and the process for notification, allocation, and dispute resolution were all still unsettled. For a class-action settlement involving authors, publishers, ownership claims, and a large pool of works, those pieces determine who is included and who gets paid.
Alsup suggested the agreement was “nowhere close to complete.” That means Anthropic and the authors' lawyers must revise the settlement process before the court will consider giving it preliminary approval.
One central concern is how payments would work when more than one person or entity claims rights in the same work. The judge wants clearer instructions requiring “anyone with copyright ownership” to opt in. Bloomberg Law reported that, under the approach Alsup wants, a work would not be covered if even one rights holder opts out.
The size of the case changed dramatically
The scale of the case is one reason the settlement is attracting scrutiny. Alsup had certified a class that included up to 7 million claimants whose works were illegally downloaded by Anthropic. The proposed settlement, however, covers less than 500,000 works.
As of this writing, Alsup said the list of covered works spans about 465,000. The Authors Guild, which consulted on the case and is part of a Working Group helping to allocate claims of $3,000 per work to authors and publishers, explained why the number fell so sharply.
According to the Authors Guild, after accounting for duplicates, foreign editions, unregistered works, and books missing class criteria, “only approximately 500,000 titles meet the definition required to be part of the class.”
That narrower list matters for authors because it determines who can participate in the settlement. It also shapes the broader debate over whether the payout is enough in a case where Anthropic could have owed more than $1 trillion in damages.
Why authors and critics are watching closely
Critics fear Anthropic may resolve the case too cheaply. The source article notes that the settlement would require Anthropic to pay a small fraction of its total valuation, currently $183 billion, and that it does not require the company to admit wrongdoing.
Defector noted that Anthropic continues raising billions based on models trained on authors' works. Most recently, Anthropic raised $13 billion in a funding round, making back about 10 times the proposed settlement amount after announcing the deal.
The Authors Guild has offered a different view of the tradeoff. It said “the Copyright Act gives courts discretion to award statutory damages of at least $750 and no more than $150,000 per infringed work when the infringement is willful, as is the case here.”
At the same time, the group said maximum damages are rare “when there are a large number of works at issue.” Authors suing Anthropic chose what the Authors Guild described as “a strong payout without the risks of trial,” rather than a legal fight that “could tie up the case for years.”
The group also said the settlement “achieved a certain, immediate result that sends a powerful signal to the industry that piracy will cost you a lot.” It suggested the deal could push more AI companies to avoid piracy and pay to license content for training.
The claims process is now the pressure point
Alsup's questions show that the court is focused not only on the dollar amount, but also on whether authors can actually understand and use the claims process. He raised concern that class members often “get the shaft” in class actions where attorneys stop caring after monetary relief is granted.
The judge also warned that a weak notification process could leave Anthropic exposed to future claimants “coming out of the woodwork later,” even after paying more than $1 billion. At the hearing, he said, “When they pay that kind of money, they’re going to get the relief in the form of a clean bill of health going forward.”
Alsup also said he felt “misled” and asked for more information about the claims process. He added, “I have an uneasy feeling about hangers-on with all this money on the table.”
To move the process forward, the judge set a schedule. Lists of covered works and class members are to be finalized by September 15. The claims process is to be finalized by September 25. That could allow the court to potentially preliminarily approve the settlement by October 10.
What authors should expect next
The Authors Guild said that once the list is finalized, likely by October 10, a searchable database will be created so authors can check whether their works are covered. Until then, authors can submit contact information through a website set up to manage the settlement process.
If the settlement is ultimately approved, the claims process will likely begin this fall. The Authors Guild said, “If your book is included in the class list, you will receive a formal notice by mail or email from the settlement administrator.”
The group said that notice will explain the settlement terms, rights, and next steps. For now, the key point is that the $1.5 billion settlement is not rejected permanently, but it has not earned the court's preliminary approval. Alsup wants the parties to finish the mechanics before authors are asked to accept the deal.