Writers Clash With Publishers and Agents Over Anthropic’s $1.5B Settlement Payouts
Faulty records and questionable claims complicate disbursements from the landmark AI copyright deal
Key highlights · 3 min read
- Authors expecting compensation from Anthropic's $1.5 billion copyright settlement are running into an unexpected hurdle: traditional publishers and literary agents are stepping in to claim their pa…
- The dispute stems from a landmark class action agreement granted final approval in July, following a legal ruling that Anthropic's unauthorized downloading of pirated books for model training breac…
- Yet many creators say that distribution framework is breaking down in practice.
The Scale ReportAuthors expecting compensation from Anthropic's $1.5 billion copyright settlement are running into an unexpected hurdle: traditional publishers and literary agents are stepping in to claim their payouts. As notification notices roll out, numerous writers report that former publishing partners and agencies are attempting to seize portions, or even the entirety, of the compensation allocated for pirated works.
The dispute stems from a landmark class action agreement granted final approval in July, following a legal ruling that Anthropic's unauthorized downloading of pirated books for model training breached copyright law. As reported by TechCrunch, the settlement covers approximately 500,000 book titles, awarding $3,000 per affected work. Under the deal, proceeds from books still actively in print are split equally between authors and publishers, while self-published writers or those whose publication rights reverted before August 10, 2022, are entitled to 100% of the funds.
Yet many creators say that distribution framework is breaking down in practice. Thriller novelist April Henry revealed that publisher HarperCollins submitted a claim for a title whose rights had reverted to her 17 years ago. Author Courtney Milan, the pen name of former law professor Heidi Bond, publicly warned against literary agencies attempting to take fees from the settlement, pointing out that agents possess no legal copyright ownership in the titles they represent.
Victoria Strauss, a representative for the industry watchdog blog Writers Beware, noted that the influx of author grievances falls into two primary buckets: publishing houses requesting the full 100% payout instead of their 50% contractual split, and companies filing claims on titles where rights have already reverted to the author. While Strauss observed that some publishers attributed the filings to administrative errors, she emphasized that the sheer volume of identical reports points to widespread, systemic administrative flaws rather than isolated technical bugs.
Mary Rasenberger, chief executive of the Authors Guild, framed the issue as an inevitable byproduct of antiquated publishing archives rather than calculated opportunism. Rasenberger indicated that inadequate cataloging across corporate databases, combined with an intricate claims mechanism, created the widespread discrepancies, though authors can formally dispute wrongful claims.
Why It Matters
The friction exposes a fundamental challenge facing large-scale AI copyright restitution. While tech companies are being forced to pay billions for training datasets, the fragmented, decades-old paperwork of legacy media companies makes fair distribution a logistical nightmare. If tracking ownership for half a million cataloged books creates this much chaos, future compensation models for visual artists, musicians, and digital creators will face even steeper hurdles.
Reporting based on coverage from AI News & Artificial Intelligence | TechCrunch.




