Disputes over who should receive compensation are emerging after Anthropic agreed to a landmark settlement with authors whose books were allegedly used to train its AI systems.
A new dispute has emerged around Anthropic’s $1.5 billion copyright settlement with authors, challenging claims made by some publishers and literary agents over a share of the compensation. The settlement stems from allegations that Anthropic used copyrighted books without permission while developing its AI models. The agreement, approved earlier this year, is expected to cover more than 482,000 works, with compensation estimated at $3,000 per eligible book.
While the settlement was intended to resolve a major copyright dispute between authors and the AI company, questions are now shifting towards how the money should be divided.
Publishers face questions over rights.
Some authors argue that publishers and agents are seeking payments for works where their contractual rights may have changed or reverted to the authors. The disagreement centers on whether publishers remain entitled to compensation for individual books and, if so, what portion they should receive. Reports cited by TechCrunch indicate that authors have objected to claims they believe go beyond the rights held by publishers or agents.
Under the settlement framework, compensation for certain books can be divided between authors and publishers depending on the rights associated with the work. This has made publishing contracts and rights ownership central to the payout process.
A wider AI copyright debate.
The dispute highlights a larger issue surrounding AI training data and copyright. Authors, publishers and technology companies are still negotiating where ownership and compensation should stand when creative works are used to develop generative AI systems.
Anthropic’s case is significant because the settlement followed allegations involving large-scale use of books in AI training. The agreement has been described as one of the largest copyright settlements involving AI.
The disagreement shows that settling with an AI company does not necessarily end the legal and financial complications around copyrighted works. For writers, publishers and agents, determining who owns the relevant rights could be as important as securing compensation. The Anthropic settlement therefore remains a significant test for the evolving relationship between authors, publishers and artificial intelligence companies, particularly as copyright disputes over AI-generated and AI-trained content continue to expand.
