Anthropic’s $1.5 billion settlement with authors and publishers is the largest known copyright settlement in US history. But the presiding judge hit the brakes at the fairness hearing on May 14.
What this is about
Authors sued in 2024, claiming Anthropic trained Claude on datasets containing hundreds of thousands of copyrighted books — downloaded from piracy sites like Library Genesis and Pirate Library Mirror. Anthropic didn’t fundamentally contest the allegations and agreed to the record-setting settlement.
What the judge wants
US District Judge Araceli Martinez-Olguin didn’t sign off at the San Francisco hearing. She’s asking for more detail on attorney fees and payments to lead plaintiffs. With a settlement this size, she wants to know exactly where the money is going.
There’s also pushback from authors themselves: some think the amount is too low, others argue the lawyers are getting too big a cut, and still others feel wrongly excluded as rights holders.
Why this matters
This case sets the benchmark for the entire AI industry. If the settlement goes through, other companies have a reference point for what copyright violations through AI training actually cost. $1.5 billion sounds like a lot — but spread across hundreds of thousands of affected works, it’s significantly less per book than many authors were hoping for.
My take
Anthropic is doing the right thing here — or at least the pragmatic thing. Instead of fighting in court for years, they’re settling the largest AI copyright case in history. That’s smart because it creates certainty and reinforces their image as the responsible AI company.
But the judge is right to scrutinize this closely. When $1.5 billion is on the table, transparency about who gets what is essential. The answer to that question will shape the next wave of AI copyright cases in a big way.
Sources: Courthouse News Service · Claims Journal · Authors Alliance