Authors Challenge Publishers, Agents Over Anthropic AI Settlement Payments

A wave of apprehension is spreading among authors who anticipated receiving their share of Anthropic's substantial $1.5 billion copyright settlement, designed to compensate creators whose works were used to train the AI model. Many writers are reporting receiving unsettling notifications indicating that publishers and even literary agents are asserting claims on payments they believe are rightfully theirs alone, sparking a contentious debate over digital rights and the intricacies of large-scale legal settlements.

Quick summary

• Authors are reporting that publishers and literary agents are making claims on their payments from Anthropic's $1.5 billion copyright settlement.
• The settlement mandates a 50-50 split between authors and traditional publishers for in-print books, with authors receiving 100% for self-published or out-of-print works where rights have reverted.
• Disputes frequently involve publishers claiming payments for books whose rights have long since reverted to the author, or seeking a full 100% share instead of the stipulated 50%.
• Literary agents, who are not designated as rightsholders in the settlement terms, are also reportedly attempting to claim percentages of the author payouts.

Why it matters

This unfolding dispute is more than a mere administrative glitch; it strikes at the heart of author compensation, intellectual property rights, and the evolving relationship between creators and the traditional publishing industry in the AI era. For individual authors, particularly those whose livelihoods depend on such income, these contested claims represent a significant financial setback and an erosion of trust. The Anthropic settlement was hailed as a landmark decision affirming creators' rights against unauthorized AI training, but the current payment controversies risk undermining its intended impact.

The situation highlights systemic challenges within the publishing world, particularly regarding archaic record-keeping practices that struggle to keep pace with digital rights management. If publishers are indeed making erroneous claims on a widespread scale, it raises questions about accountability and the transparency of rights management, potentially necessitating clearer industry standards and better infrastructure for tracking intellectual property. Moreover, the involvement of literary agents in claiming portions of payments for which they are not rightsholders introduces a new layer of complexity, challenging established norms of representation and compensation.

Beyond the immediate financial implications for authors, this scenario offers a crucial precedent for future AI-related copyright settlements. The difficulties encountered here will likely inform how similar disputes are structured and managed, pushing for more robust mechanisms to ensure fair and accurate distribution of funds to original creators. It underscores the urgent need for a unified approach to digital rights in an increasingly automated content landscape, ensuring that the benefits of technological advancement do not disproportionately accrue to intermediaries at the expense of those who produce the original work.

Background

The genesis of this issue lies in a significant legal battle over artificial intelligence training data. Anthropic, a prominent AI development company, found itself at the center of a class-action copyright lawsuit. The core of the legal challenge revolved around the use of copyrighted material to train its large language models without explicit permission or compensation to the creators.

A pivotal moment in the litigation occurred when a judge issued a ruling that distinguished between two key aspects: while training AI models on copyrighted material was deemed legal under fair use doctrine, the act of *pirating* that material to facilitate such training was not. This nuanced judgment set the stage for a settlement, ultimately culminating in Anthropic agreeing to pay $1.5 billion to resolve the class-action suit. The settlement received its final judicial approval in July of the preceding year, clearing the path for the distribution of funds to affected authors.

Under the agreed-upon terms, the settlement was designed to compensate authors of nearly 500,000 distinct titles identified as having been improperly used. Each pirated work was set to receive a payment of $3,000. Crucially, the settlement outlined a specific distribution mechanism: for books still in print with a traditional publisher, the payment was to be split equally, 50-50, between the author and the publisher. However, if a book was self-published, or if its rights had reverted entirely to the author because the publisher had allowed it to go out of print, the author was entitled to the full 100% of the payment. It is precisely these stipulations, particularly concerning rights reversion, that are now at the heart of the ongoing disputes.

Qnews24h insight

The current turbulence surrounding the Anthropic settlement payments reveals a deeper, more systemic challenge within the traditional publishing ecosystem: the chasm between legacy administrative practices and the rapid evolution of digital rights in the AI age. While immediate reactions might point to opportunistic 'grabs' by publishers or agents, the more plausible explanation, as suggested by industry veterans like Victoria Strauss of Writers Beware and Authors Guild CEO Mary Rasenberger, points to widespread poor record-keeping and a settlement process inherently complex.

This situation underscores a critical insight: the digital age has fundamentally altered the lifecycle of intellectual property, but the mechanisms for tracking and managing these rights have not kept pace. Many traditional publishing houses operate with antiquated systems not designed to meticulously log the precise dates of rights reversions for hundreds of thousands of titles over decades. The sheer volume and granularity of data required for a settlement of this magnitude, coupled with the varied contractual terms that exist across different authors and publishers, create an environment ripe for inaccuracies.

The 'download date' of August 10, 2022, a critical cutoff for determining rights reversion claims, exemplifies the technical intricacies that can trip up even well-intentioned parties. This isn't just about mistakes; it's about the inherent difficulty of applying a rigid, singular set of rules to a mosaic of historical, diverse contractual relationships. This scenario serves as a stark warning and a call to action for the entire creative industry: robust, transparent, and digitally-native rights management systems are no longer a luxury but a necessity to protect creators' interests in an increasingly AI-driven content economy. Without them, future settlements, however well-intentioned, will likely face similar, if not greater, distributive chaos.

The Disputed Claims and Industry Responses

Reports from authors across social media platforms paint a consistent picture of contested claims. Mystery and thriller writer April Henry publicly questioned HarperCollins, alleging the publisher claimed a book that had reverted to her more than 17 years prior. Simultaneously, she noted a credit alert indicating HarperCollins had been added as her employer, despite never having been in such a relationship.

Victoria Strauss, who runs the popular blog Writers Beware, has cataloged numerous author complaints. These largely fall into two categories: publishers asserting claims on works whose rights have already reverted to the author, and publishers demanding a full 100% payment when the settlement terms entitle them to only 50%. Strauss, while acknowledging the potential for errors, described the volume and uniformity of these reports as suggesting something “much more widespread and systemic” than routine glitches.

Echoing this sentiment, Mary Rasenberger, CEO of the Authors Guild, clarified that she does not perceive these issues as deliberate attempts by publishers to exploit authors. Instead, she attributes the problems to what she calls the “predictable result of bad record-keeping and a confusing settlement process.” Some publishers have already begun to acknowledge these errors, reportedly asking Anthropic to rectify incorrect claims.

Unexpected Agent Involvement

Compounding the frustration, authors are also reporting that literary agents are making claims on parts of their settlement payments. Strauss expressed surprise at this development, emphasizing that agents are not typically considered rightsholders in the books they represent. Author Courtney Milan, a former law clerk and law professor, took to Bluesky to voice her strong disapproval, stating, “Apparently some agents are trying to claim percentages on the Anthropic settlement, and I do not REMOTELY think they should do this, what the f***, stop that sh**!”

This situation adds another layer of complexity, raising questions about the scope of agent-author agreements and whether existing contracts implicitly or explicitly allow for such claims on settlement funds that are ostensibly for copyright infringement.

Recourse and Navigating the Process

In response to the growing number of disputes, both Courtney Milan and the Authors Guild have taken steps to provide authors with guidance on how to challenge incorrect payment allocations. A critical factor in disputing claims, particularly regarding rights reversion, is the specific date. For an author to legitimately claim 100% of a payment for a reverted book, the rights reversion must have occurred prior to August 10, 2022. This date is crucial as it represents the 'download date' established within the settlement agreement for identifying pirated works.

Authors are advised to meticulously review their contracts and any reversion letters to ascertain the exact status of their book rights. The process of challenging these claims requires diligence and a clear understanding of the settlement's specific parameters. This situation underscores the importance for creators to maintain accurate records of their intellectual property rights and to understand the legal nuances of such complex settlements.

Sources

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

FAQ

• What is the Anthropic copyright settlement?
  The Anthropic copyright settlement is a $1.5 billion agreement reached after a class-action lawsuit. It compensates authors whose copyrighted works were used without permission to train Anthropic's AI models, based on a ruling that pirating material for AI training is illegal, even if fair use applies to the training itself.
• Why are authors disputing payments from the settlement?
  Authors are disputing payments because publishers are allegedly claiming shares for books whose rights have reverted entirely to the author, or attempting to take 100% of payments when only a 50% split is stipulated for in-print books. Additionally, some literary agents, who are not rightsholders, are also reportedly making claims on author payments.
• How is the settlement money supposed to be split?
  For books still in print with a traditional publisher, the settlement money is to be split 50-50 between the author and the publisher. If a book was self-published or its rights had reverted to the author (meaning it was out of print and the author regained full control), the author is entitled to 100% of the payment.
• What should authors do if their payment is being incorrectly claimed?
  Authors should gather evidence of their rights, particularly any documentation proving rights reversion occurred before August 10, 2022 (the settlement's 'download date'). Organizations like the Authors Guild and legal experts are providing guidance on how to dispute these claims through the official settlement channels.