Anthropic’s $1.5 Billion Settlement Has Exposed the Messy Economics of AI and Books 

Who gets the money when a book used to train an AI chatbot is claimed by both its author and publisher?

The question is becoming increasingly important as thousands of authors await payments from Anthropic’s $1.5 billion copyright settlement, the largest such settlement in U.S. history, while disputes emerge over who owns the rights to books covered by the agreement.

Anthropic, the artificial intelligence company behind the Claude chatbot, agreed to pay the money to resolve claims that it used pirated copies of books to train its AI systems. Under the settlement, rights holders could receive roughly $3,000 for each qualifying book that was used.

But some authors are now concerned that publishers could seek a share of the settlement payments, potentially reducing the amount that reaches writers whose books were included in the case.

The issue has surfaced as a settlement administrator works to determine who holds the rights to some of the more than 482,000 books covered by the agreement in the class-action lawsuit. The settlement was approved in July by a judge in the U.S. District Court for the Northern District of California.

For many of the books, ownership is undisputed. But in a number of cases, authors and publishers have submitted competing claims to the settlement administrator, which recently began notifying parties when discrepancies were identified, said Mary Rasenberger, chief executive of the Authors Guild, a professional association that says it has more than 18,000 members.

In an interview, Rasenberger said the settlement process was ripe for such misunderstandings.

Publishing contracts often divide rights between authors and publishers in complicated ways, particularly for older books that may have passed through multiple publishing arrangements. Determining who is entitled to compensation can therefore become difficult when a settlement covers hundreds of thousands of titles.

The dispute comes at a sensitive time for authors, many of whom see the Anthropic settlement as an important recognition that their copyrighted work has commercial value in the development of generative AI.

AI companies have increasingly faced lawsuits from writers, news organizations, artists and other copyright holders who argue that their work has been copied or used to train AI systems without permission.

Anthropic has maintained that its use of books for AI training was protected by copyright law in at least some circumstances, but the litigation has produced a series of legal challenges over how existing copyright principles apply to generative AI.

The settlement with authors and other rights holders does not resolve the broader legal debate over whether AI companies can use copyrighted material to train their models.

Instead, it provides a mechanism for compensating rights holders whose works were allegedly included in Anthropic’s training materials.

The $1.5 billion figure has attracted particular attention because of its size. It reflects the scale of the potential financial exposure facing AI companies as they seek to build increasingly capable models using vast amounts of data.

For authors, however, the headline figure does not necessarily translate into an equal payment for every writer whose work is covered.

The settlement administrator must first determine which books qualify and who has the legal right to receive the money. That process could become more complicated when authors and publishers both assert ownership interests.

Rasenberger said authors who believe publishers are improperly claiming rights could challenge those claims through the settlement process.

If disputes cannot be resolved, a court-appointed arbitrator will make the final determination, Rasenberger said. The Authors Guild, she added, “will fight tooth and nail for them if they are, in fact, being screwed over.”

The dispute illustrates how copyright ownership can become complicated when a work changes hands or when contractual arrangements give different parties rights over different uses of a book.

A publisher may have contractual rights to publish and distribute a work, for example, while an author may retain other rights. The precise terms can vary significantly from one publishing agreement to another.

Those distinctions matter because the settlement is tied to rights in books that Anthropic allegedly obtained and used for AI training.

For authors, the concern is not simply how much money Anthropic is paying. It is also whether the payment will ultimately go to the person or entity that has the appropriate legal claim.

The issue could affect thousands of authors whose works appear in the settlement database.

The scale of the case means even a relatively small number of contested claims could create a significant administrative burden. The settlement administrator must reconcile claims involving hundreds of thousands of individual works while giving authors and publishers an opportunity to contest the ownership information.

The process also highlights the difficulty of applying traditional publishing contracts to a new technological use of copyrighted material.

AI training was not contemplated by many older publishing agreements, leaving authors and publishers to determine how existing contractual language applies to compensation arising from the use of books in AI systems.

The Anthropic case is part of a wider confrontation between the technology industry and creative professionals over the use of copyrighted material to develop AI.

Generative AI companies need enormous quantities of text, images, code and other information to train their models. Publishers and creators, meanwhile, argue that the commercial use of their work should require authorization and compensation.

The resulting lawsuits could help establish rules for a rapidly expanding industry, but settlements such as Anthropic’s also provide a way to resolve individual disputes without waiting years for courts to settle every question of copyright law.

For publishers, the settlement presents an opportunity to recover money connected to works over which they may hold contractual rights. For authors, it represents compensation for the use of creative work that they say was taken without permission.

That difference in perspective is now creating a second dispute inside the settlement itself.

The question of who owns a book for settlement purposes could ultimately be as important to some authors as the size of the settlement.

The court-appointed arbitrator’s decisions could therefore help establish how competing claims between authors and publishers are handled, although those rulings would be tied to the settlement rather than necessarily establishing a broad new rule for the publishing industry.

The outcome could also influence how authors and publishers negotiate future contracts as AI companies seek licenses for books and other copyrighted material.

For authors, the Anthropic case has already underscored the economic value of their work in the AI era. The emerging disputes over the settlement are now forcing another question: who should benefit when that value is converted into a financial payment?

Stay ahead of the Stories shaping our world. Subscribe to Impact Newswire and join our 
WhatsApp Channel for updates on global tech, business, and innovation—all in one place.

Dive deeper into the future with the Cause Effect 4.0 Podcast, where we explore the ideas, trends, and technologies driving the global AI conversation.

Got a story to share? Contact Us to reach a global audience with Impact Newswire.