Judge approves Anthropic’s $1.5 billion copyright settlement with authors - Ars Technica

What the report says

Ars Technica reported that a federal judge on Monday approved Anthropic’s $1.5 billion settlement with authors, resolving what Ars described as the largest certified copyright class action and the biggest copyright settlement of its kind. The case followed a split court finding: Anthropic’s use of books to train AI was deemed fair use, while its alleged acquisition of pirated copies remained a likely copyright problem.

US District Judge Araceli Martínez-Olguín rejected most objections from authors who said the settlement undervalued their claims or that legal fees were excessive. According to Ars, the court noted that about 95 percent of the class received notice and roughly 91 percent of affected authors and publishers had submitted claims. The expected payment is around $3,000 per work, which the judge said exceeded minimum statutory damages by a multiple of four. Only 350 class members opted out, while dozens of objections or late opt-out efforts were denied.

The judge approved the overall settlement but reduced compensation for lawyers and class representatives. Attorneys had initially sought $300 million, later lowering the request to about $187 million; the court cut fees to roughly $101 million and required later accounting that could further reduce the amount. Awards for the three lead author plaintiffs were lowered from the requested $50,000 each to $15,000.

Ars also reported that Anthropic opposed last-minute opt-outs, including one from author and entertainment lawyer Donald Passman. The judge allowed only two late requests after finding excusable neglect. The decision matters beyond this case because it leaves intact a major AI fair-use ruling while imposing a large settlement tied to alleged piracy of copyrighted books.