Lawyers for the authors hailed the deal as unprecedented. Justin Nelson, one of the lead attorneys, said: 'As best as we can tell, it's the largest copyright recovery ever, and the first of its kind in the AI era.'
The lawsuit alleged that Anthropic downloaded nearly 500,000 copyrighted works without permission, many from shadow libraries such as LibGen, to train Claude. The plaintiffs included well-known writers like Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson.
In June, a federal judge ruled that while training AI models on copyrighted text could, under some circumstances, qualify as 'fair use,' the act of obtaining pirated copies remained illegal. That ruling opened the door for the authors' case to move forward, according to court records cited by AP.
Facing the possibility of billions in damages at trial this December, Anthropic opted to settle. Aparna Sridhar, the company's deputy general counsel, said in a statement: 'This agreement resolves the legacy claims brought by the plaintiffs. We remain committed to building safe AI systems that benefit society.'
The settlement requires Anthropic not only to compensate authors but also to destroy the pirated datasets it had acquired.
The case is being closely watched across the tech sector. Dozens of AI developers, including OpenAI, Microsoft, and Meta, face similar lawsuits over the use of copyrighted material to train large language models.
Rania Umutoni
