Anthropic’s $1.5 billion payout closes a piracy case but not AI’s copyright fight

Anthropic’s $1.5 billion payout closes a piracy case but not AI’s copyright fight
Anthropic has won final approval for a $1.5 billion settlement with authors and publishers, ending one of the biggest book-piracy cases tied to AI training. But the ruling resolves a narrow dispute over how books were obtained, not the larger legal battle over whether AI companies can train models on copyrighted works without permission.

The settlement is being described as the largest copyright recovery on record, with payments of about $3,000 per work for covered books. Judge Araceli Martínez-Olguín said the deal would provide “meaningful relief,” according to reporting on the final order. Anthropic, for its part, said it was “pleased that more than 91% of authors and publishers covered by the settlement have claimed their share of the payment” and that it was “looking forward to bringing this matter to a close.”

Supporters of the agreement point to that participation rate as evidence that most claimants saw the deal as worthwhile. Ars Technica reported that only 350 class members opted out, while the judge also trimmed attorneys’ fees to about $101 million, well below earlier requests. In the court’s view, the roughly $3,000-per-work payment was also several times higher than minimum statutory damages.

But critics say the headline number masks a more limited outcome. Some authors argued the per-book payout was too low and objected that the case would not produce a definitive appellate ruling on AI training itself. That distinction matters because the earlier court rulings suggested Anthropic could claim fair use for training on legally acquired books, while still facing liability for what one report called the “Napster-style downloading of millions of works.”

In other words, this case punishes piracy, not necessarily AI training as a practice. As TechCrunch noted, the final approval “settles one case, but it doesn’t resolve the broader issue of using copyrighted works to train AI models.” That leaves similar lawsuits against other AI companies very much alive.

Continue reading https://foxvector.com/stories/019f9393-9b47-21b5-721e-0b543c7de1ad