Some of those fees include future work that lawyers will do to distribute funds, and some authors argued that lawyers’ projections for time spent were alarmingly inflated. Acknowledging these concerns, Martínez-Olguín wrote in a footnote that the court decided to take “an additional step to protect the interests of the Class.” Once payouts have been finalized, lawyers will be required to file a post-distribution accounting, and if their fees are less, then the court “may reduce” them, Martínez-Olguín said.

Although the settlement likely relieves the three authors who spent years defending their works and 506,194 works from potential class members, the lead plaintiffs are likely disappointed by a decision to reduce their requested service awards from $50,000 to $15,000.

But Martínez-Olguín said that the higher amount was “unreasonable,” despite authors spending substantial time and resources on both litigation and the settlement negotiation. Without a sign that authors may face retaliation for filing the lawsuit, the judge ruled that lower awards were due.

In a statement to Reuters, lead plaintiffs said that the settlement “brings us one step closer to real accountability for Anthropic and puts all AI companies on notice they can’t shortcut the law or override creators’ rights.”

It’s unlikely, but still possible, that authors may receive more than $3,000 per work.

“If any funds remain in the Settlement Fund after all Valid Claims are paid, the Parties anticipate a redistribution of the remaining funds to Settlement Class Members unless it is economically infeasible to do so,” Martínez-Olguín said.

Authors’ lawyers celebrated the judge’s approval, writing in a press release that “rightsholders came together to call foul” on Anthropic’s piracy and describing support for the settlement among them as “overwhelming.”

Anthropic also seems happy to be done arguing about the settlement. In a statement provided to Ars, Anthropic’s deputy general counsel, Aparna Sridhar, said that Anthropic is glad the case’s landmark ruling established that its AI training was fair use and is happy that most rightsholders have already filed claims.