Authors push back as publishers and agents seek share of Anthropic settlement

Authors are encountering unexpected claims on their share of the $1.5 billion Anthropic copyright settlement, with publishers and literary agents attempting to secure portions of payments intended for writers. The settlement, which compensates authors for AI models trained on copyrighted works, stipulates a 50-50 split between author and publisher for in-print books, with authors receiving the full amount for reverted or self-published titles. However, numerous authors report publishers erroneously claiming full payment for books with rights that reverted decades ago, or attempting to take the entire 50% share when only entitled to half. Literary agents are also reportedly making claims despite not being rights holders. While some suggest these issues stem from poor record-keeping and a complex settlement process, the widespread and repetitive nature of these errors has led some to suspect a more systemic problem. Authors are now actively disputing these claims, navigating specific deadlines and procedures to ensure they receive their rightful compensation.

AI Signal Decode

The core issue revolves around the distribution of funds from Anthropic's $1.5 billion copyright settlement, which aims to compensate authors whose works were used to train AI models. The settlement terms, finalized in July, grant authors $3,000 per pirated work, with a 50-50 split for books still under traditional publishing contracts and 100% for authors of self-published or out-of-print works whose rights have reverted. This structure has created a conflict as publishers and literary agents appear to be overreaching in their claims, with authors reporting instances where publishers are demanding full payment for books whose rights reverted years ago or claiming more than their entitled 50% share.

The market implications are significant, as this dispute could set a precedent for how future AI copyright settlements are handled and how publishing contracts are interpreted in the context of AI training data. It highlights a potential for exploitation within the existing legal and contractual frameworks, particularly as publishers and agents navigate the complexities of digital rights management and AI-generated content. The controversy also underscores the ongoing tension between creators and entities utilizing their work for AI development, potentially leading to increased scrutiny and demand for clearer contractual language and more transparent settlement processes.

From a technical and legal standpoint, the situation exposes the challenges in accurately tracking and attributing rights for a vast number of literary works, especially those with complex ownership histories involving rights reversions. The settlement's reliance on specific dates for rights reversion (e.g., August 10, 2022, for the download date) introduces a layer of complexity that may be contributing to errors in payment allocation. What to watch next includes how Anthropic and the involved publishers address these systemic complaints, whether regulatory bodies step in, and if authors organize further to protect their interests, potentially leading to further legal challenges or revised industry practices for AI-related copyright compensation.