Authors push back as publishers and agents make claims on Anthropic settlement

Authors Push Back on Publishers and Agents Over Anthropic AI Settlement Shares

The world of artificial intelligence continues to evolve at a dizzying pace, bringing with it not just technological breakthroughs but also complex legal and ethical challenges. One of the most significant recent developments in the AI landscape is the landmark $1.5 billion copyright infringement settlement by Anthropic, the company behind the Claude family of large language models. However, as the dust settles and payouts begin, a new conflict is emerging: authors are increasingly vocal about what they perceive as publishers and literary agents claiming more than their fair share of these settlement payments. This dispute highlights the ongoing tension between creators, traditional publishing powerhouses, and the burgeoning AI industry, all grappling with how to fairly compensate for intellectual property in the digital age.

The Genesis of the Anthropic Settlement

Anthropic, an American AI safety company founded in 2021 by former OpenAI researchers, is known for developing its Claude AI models. These large language models (LLMs) are trained on vast amounts of text data to generate human-like responses. The core of the legal challenge against Anthropic, known as Bartz v. Anthropic, stemmed from allegations that the company downloaded millions of copyrighted books from "shadow libraries" like Library Genesis (LibGen) and Pirate Library Mirror (PiLiMi) to train its AI systems without permission. Court documents revealed that Anthropic downloaded over 7 million books from these pirated sites. The lawsuit, initially filed in August 2024 by authors Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson, accused Anthropic of large-scale copyright infringement. While a U.S. District Court judge ruled in June 2025 that using legally acquired books for AI training could be considered fair use, the judge denied Anthropic's request for summary judgment regarding piracy, finding that the use of pirated material was not fair use. This distinction was crucial, paving the way for the settlement. In August 2025, Anthropic agreed to a historic $1.5 billion settlement, believed to be the largest copyright recovery in U.S. history and the first of its kind in the AI era. This settlement received final approval from Judge Araceli Martinez-Olguin of the U.S. District Court for the Northern District of California on July 20, 2026. The agreement covers approximately 500,000 titles that were downloaded from pirated sources and registered with the U.S. Copyright Office. Each qualifying title is set to receive approximately $3,000 in compensation.

The Heart of the Dispute: Who Owns the AI-Generated Rights?

The settlement was intended to provide compensation to rightsholders whose works were used without permission. However, as the $1.5 billion begins to be distributed, authors are discovering that their publishers and, in some cases, literary agents, are making claims on these payments. This has ignited a fresh debate about the interpretation of existing contracts—many of which were drafted long before the advent of large language models and the concept of AI training data. The core issue revolves around who truly owns the "AI rights" or, more precisely, the rights to compensation for the unauthorized use of copyrighted material for AI training. The settlement agreement outlines a default split for trade and university press contracts of 50/50 between publishers and authors, reflecting industry norms. However, for authors who are the sole rightsholder—such as self-published authors or those whose rights have reverted to them—they are entitled to 100% of the award.

Authors' Arguments: A Fair Share or an Overreach?

Many authors feel that publishers and agents are making opportunistic claims on funds that should rightfully belong solely to the creators. Their arguments often center on several key points:

• Reverted Rights: A significant point of contention arises when an author's rights to a book have reverted from the publisher back to the author. The Authors Guild states that if an author's rights reverted before August 10, 2022 (the "download" date in the settlement), they should be entitled to 100% of the funds for that title. Authors are reporting instances where publishers are still claiming a share for books whose rights have clearly reverted, sometimes even demanding 100% of the payment.
• Nature of the Infringement: Authors argue that the infringement by Anthropic was directly related to the content of their creative work being used as training data. Therefore, the compensation should primarily go to the creator of that content.
• Contracts Predating AI: Many publishing contracts were signed decades ago, long before AI training data was a consideration. Authors contend that these contracts could not have possibly granted publishers rights to compensation for AI-related uses, as such uses were unforeseen.
• Agents as Non-Rightsholders: Authors have also reported literary agencies making claims for 15% to 25% of the allocation on books they represented. The general consensus, including the Authors Guild's position, is that agents are not rightsholders in the books they sell and therefore should not be claiming a portion of the settlement.

Publishers' and Agents' Counterarguments (Implied)

While explicit public statements from publishers specifically defending their claims on the Anthropic settlement are less common, their stance generally relies on interpretations of existing contracts and the nature of their partnership with authors.

• Contractual Language: Publishers may argue that broad language in older contracts covering "all forms of exploitation" or "subsidiary rights" could be interpreted to include new technologies like AI training data.
• Investment and Risk: Publishers invest in editing, marketing, distribution, and legal protection for books. They might argue that their contribution to bringing a work to market entitles them to a share of any compensation related to that work, even for unforeseen uses.
• Administrative Role: Publishers often handle copyright registration and manage intellectual property on behalf of authors, which they may see as justifying a share of such settlements.
• Default Splits: The settlement administrator has provided a default 50/50 split between authors and publishers for traditionally published works, which some publishers may simply adhere to, assuming it's the correct and agreed-upon division.

Navigating the Legal Labyrinth and Industry Implications

The conflicting claims have led to a complex situation, with the settlement administrator now informing both sides when competing claims exist for the same titles. Authors are being advised by organizations like the Authors Guild to actively contest incorrect claims. This often involves providing documentation, such as reversion letters, to prove their sole ownership. Some publishers have already acknowledged mistakenly selecting a 100% allocation and are updating their claims. This dispute is more than just about money; it's about setting precedents for how intellectual property is valued and compensated in the age of AI. The outcome of these internal battles could significantly influence future publishing contracts, leading to more explicit clauses regarding AI training data and digital rights. It also underscores the importance of authors understanding their contracts and actively managing their intellectual property rights. The Anthropic settlement, while historic in its scale, only covers past acquisition and copying of pirated works up to August 25, 2025. It does not give Anthropic future rights to use pirated books, nor does it resolve claims related to AI outputs or future conduct. This means the legal challenges and the ongoing dialogue about AI and copyright are far from over. Other major lawsuits, such as those against OpenAI and Microsoft's Copilot, are still pending, and the music industry is also pursuing legal action against Anthropic for alleged misuse of song lyrics.

Looking Ahead

The pushback from authors against publishers and agents over the Anthropic settlement payments is a clear signal that the creative community is increasingly assertive about its rights in the face of generative AI. This situation will likely lead to: Rethinking Contracts: Expect to see new clauses in publishing and agency agreements that explicitly address AI training, data usage, and compensation for such uses. Greater Transparency: There will be increased demand for transparency from AI companies regarding their training data sources and from publishers/agents regarding how they manage and claim rights in the AI context.

• Continued Legal Scrutiny: The legal landscape around AI and copyright is still nascent, and more lawsuits and settlements are anticipated as the industry matures. Ultimately, this ongoing saga highlights the critical need for clear, equitable frameworks that respect creators' rights while fostering innovation in AI. The Anthropic settlement was a significant step, but the complexities of distributing those funds reveal deeper issues that the creative and tech industries must collectively address.

Frequently Asked Questions

What is the Anthropic settlement about?

The Anthropic settlement is a landmark $1.5 billion agreement reached by AI company Anthropic to resolve a class-action copyright infringement lawsuit, Bartz v. Anthropic. The lawsuit alleged that Anthropic unlawfully acquired and used millions of pirated books from "shadow libraries" to train its Claude large language models.

Who is eligible to receive money from the Anthropic settlement?

Rightsholders of books that Anthropic downloaded from LibGen and PiLiMi, provided the books were registered with the U.S. Copyright Office in a timely manner and have ISBN or ASIN numbers, are eligible. This includes both authors and publishers.

Why are authors pushing back against publishers and agents regarding the settlement?

Authors are pushing back because they believe publishers and literary agents are claiming more than their rightful share of the $1.5 billion settlement. This is especially contentious when authors' rights to a book have reverted to them, or when agents, who are not rightsholders, attempt to claim a percentage.

What should authors do if they believe a publisher or agent is making an incorrect claim on their Anthropic settlement payment?

Authors should immediately contact the settlement administrator to report the discrepancy. The Authors Guild advises authors to submit documentation, such as reversion letters, to prove their sole ownership of rights for a particular title.