In the long shadow of a colossal copyright clash, the legal teams who wrangled a $1.5 billion settlement with Anthropic have stepped forward with a price tag of their own: a cool $300 million.
The request landed in a California federal court this week, courtesy of the firms steering the class action—a duo that now argues their cut, exactly 20% of the settlement fund, is modest by the standards of mega-cases. They say the battle demanded more than 26,000 hours, a trek they describe as “radiating with risk.”
This saga began when authors discovered their works—hundreds of thousands of them—had allegedly been absorbed into the training diets of Anthropic’s AI models. The company chose to settle in October, agreeing not just to pay out, but to wipe the disputed datasets entirely and swear they wouldn’t power its commercial systems ever again.
For writers caught in the swirl, the deal promises over $3,000 per copyrighted work—an unprecedented payout in a copyright class action. Yet some still aren’t satisfied. Authors have until mid-January to bow out and pursue their own claims, and a long spring hearing awaits, where the judge will sift through praise, protest, and the intricate architecture of the agreement.
Anthropic maintains it did nothing wrong, even as it embraced the settlement. The company now has the option to push back against the attorneys’ fee haul.
In April, the court will take up the final question: Is this sweeping settlement—built on vanishing datasets, historic payouts, and a hefty fee request—a fair closing chapter to the AI-era book dispute? Only then will the story truly end.
