The Battle Over Anthropic's Copyright Settlement: Authors Push Back Against Publishers and Agents Claiming a Cut
Authors fight back as publishers and agents try to claim shares of Anthropic's copyright settlement funds.
After Anthropic reached a copyright infringement settlement with authors, publishers and literary agents invoked existing contract clauses to claim a portion of the funds — sparking fierce opposition from the author community. Authors argue that since AI models consumed their creative work directly, compensation should flow to creators, not distribution intermediaries. The dispute stems from severely outdated publishing contracts drafted before the rise of generative AI, whose broad "subsidiary rights" language publishers are now stretching to cover entirely new legal territory. The case is prompting calls across the industry to explicitly address AI-related rights in future contracts.
The Power Struggle Behind the Copyright Settlement
The copyright dispute settlement between Anthropic and a group of authors was supposed to be a milestone moment for creators seeking compensation. But the situation has unfolded in ways many didn't anticipate — publishers and literary agents have stepped in, each looking to claim a share of the settlement funds. This has sparked a fierce backlash from the author community and thrust the deeper industry tensions underlying AI copyright disputes into the public spotlight.
From a legal standpoint, damages in copyright infringement lawsuits are typically directed at the copyright holder. Copyright ownership in the publishing industry is complex: some authors retain full rights, while others have transferred some or all rights to publishers when signing contracts. This ambiguity is precisely what gives publishers legal grounds to intervene — and what planted the seeds of this conflict.
The Logic Behind Publisher and Agent Involvement
A Double-Edged Sword in Contract Terms
The core basis for publishers claiming a share lies in "subsidiary rights" or ancillary revenue-sharing clauses typically included in publishing contracts. Originally designed to cover derivative income such as licensing and adaptation rights, these clauses are now being extended to cover settlements in AI training data copyright disputes — a legal interpretation that remains highly contentious.
Literary agents present a different situation. Agents typically collect commissions of 10% to 15% under representation agreements. Some agents argue they played a substantive role in facilitating copyright protection and helping authors participate in class action suits, and are therefore entitled to compensation. Critics counter, however, that extending routine agency service fees into the realm of legal settlements is an overreach of contractual rights.
What Authors Are Actually Demanding
The authors pushing back broadly agree that publishers' and agents' claims have gone beyond what's reasonable. Their arguments center on three main points: first, the direct victims of AI companies' infringement are the authors who created the content, not the distribution channels; second, settlement funds are more akin to compensation for creative labor than traditional licensing revenue; and third, if intermediaries are allowed to siphon off large portions, the actual amount reaching authors will be drastically reduced, effectively hollowing out the settlement's purpose.
The Deeper Industry Tensions in AI Copyright Disputes
Re-Examining the Creative Value Chain
This fight over settlement funds reflects longstanding tensions in the content creation value chain of the digital era. Under the traditional publishing model, publishers shoulder significant costs for distribution, marketing, and editorial work — sharing in royalty revenue makes sense in that context. But in the specific case of AI training data copyright, publishers contributed virtually nothing. What AI models "consumed" was the authors' words and ideas, not publishers' distribution services.
Some copyright law scholars share this view: damages for AI training data infringement are closer in nature to compensation at the level of "moral rights" or personal creative authorship, and should be prioritized for the actual content creators. If publishing contracts don't explicitly cover such novel scenarios, publishers' claims to a share may have a difficult time holding up legally.
The Lag in Industry Rules and the Push to Rebuild Them
The current dispute also exposes serious gaps in publishing contract standards. The vast majority of contracts between authors and publishers were drafted before the rise of generative AI at scale. The contract language simply couldn't anticipate the entirely new risk scenario of AI training data copyright — leaving both parties able to interpret the same contract in their own favor.
This dispute is driving industry conversations about how future contracts should address AI-related rights. A growing number of author advocacy organizations are calling for new contracts to explicitly spell out the terms for AI training data usage rights, the allocation of related licensing revenue, and the distribution rules for litigation settlement proceeds — all to prevent similar disputes from recurring.
Broader Industry Implications
Lessons for AI Companies
The aftermath of Anthropic's settlement carries lessons for other AI companies facing similar lawsuits. It reveals a practical reality: even when an AI company is willing to accept liability through a settlement, the actual flow of funds can still trigger new disputes. Settlement structures need to more carefully account for the copyright ownership chain — and in some cases, it may be necessary to negotiate directly with authors, bypassing traditional publishing intermediaries.
Structural Challenges for the Creator Economy
From a broader perspective, this dispute is the latest manifestation in the AI era of the structural tension between the "creator economy" and the "platform/intermediary economy." Regardless of the ultimate legal outcome, it serves as a reminder to the entire industry: as AI reshapes how content is created and distributed, ensuring that actual creators receive fair compensation will be a core challenge requiring sustained attention and systemic solutions.
Several author advocacy organizations have already indicated they will closely monitor developments and are considering providing legal assistance to challenge publishers' claims. This battle over settlement funds is far from over.
Related articles
Catalyst: A Vision for an Enzyme-Like Testing Framework for AI Agents
A developer shared Catalyst on Reddit, an Enzyme-inspired framework for AI Agents, exploring why agents need observable, testable dev tools and the design philosophy behind them.
The Real Capability of AI Coding Agents: Best Models Complete Only 35% of Feature Development Tasks
The 'Agents on Rails' benchmark finds top AI models complete only 35% of feature development tasks. What this means for coding agents and developer teams.
How to Prevent Duplicate Refunds After an AI Agent Crashes: CellaFlow's Durable Execution Approach
How can AI agents avoid duplicate refunds after a crash without deadlocking workflows? CellaFlow uses durable execution, shared work identity, leases, and fencing to solve safety and liveness in multi-agent systems.