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

Settlement Payouts Spark Dispute

Authors expecting payments from Anthropic's $1.5 billion copyright class-action settlement are reporting that publishers and literary agents are making improper claims on their funds. The dispute centers on the distribution of the $3,000 per-work payments intended for creators whose books were pirated to train AI models. This situation highlights a critical operational breakdown in compensating rightsholders, raising questions about the feasibility of executing large-scale AI-related copyright agreements fairly and accurately.

The Details of the Discrepancy

Under the settlement's terms, payments are split 50-50 between authors and publishers for books still in print, while authors are entitled to the full amount if rights were reverted prior to August 10, 2022. However, many authors have stated that former publishers, such as HarperCollins, are claiming a share of works for which they no longer hold the rights. Observers like Victoria Strauss of Writers Beware note that the errors are widespread and systemic, not isolated glitches.

• Publishers claiming 50% for books with reverted rights.
• Publishers claiming 100% of the payment instead of the 50% share.
• Literary agencies, who are not rightsholders, also making claims on funds.

Impact on AI and Copyright Precedent

While organizations like the Authors Guild suggest the problems stem from poor record-keeping rather than malicious intent, the sheer volume of complaints points to a significant flaw in how intellectual property rights are tracked. This episode serves as a case study for the AI industry, demonstrating that even after a legal settlement is reached, the administrative infrastructure for compensating creators is not yet robust enough to handle the complexity. It underscores the deep operational challenges that accompany the legal battles over AI training data, potentially complicating future agreements between AI labs and content creators.

The logistical chaos surrounding the Anthropic settlement reveals a critical vulnerability for the AI industry: while legal frameworks for copyright are being hammered out, the back-end administrative systems for tracking and compensating rightsholders are not prepared for the scale of these agreements, creating significant reputational and operational risk.