A $1.5 billion settlement against AI giant Anthropic has been approved by a US judge.
Authors and publishers are set for payouts after judge Araceli Martínez-Olguín certified the largest ever copyright lawsuit.
However the settlement recognised that Anthropic, maker of generative AI chatbot Claude, is free to train its models on works it has legally acquired as this was judged to be fair use.
The class action of Bartz et al vs. Anthropic, filed in 2024, pertained to Anthropic training its AI models on content it had not purchased.
Writers accused Anthropic of illegally downloading and storing millions of copyrighted books from pirate sites such as Library Genesis and Pirate Library Mirror.
Judge William Alsup of the U.S. District Court for the Northern District of California issued a preliminary approval of the settlement last year
He has since retired. Now judge Martínez-Olguín has approved the payout of $3,000 per work across an estimated 500,000 works, shared among the authors and publishers who hold the rights.
Martínez-Olguín also reduced both lawyers’ fees and awards for the three lead plaintiffs.
Lawyers had originally requested 20% – $300 million – in fees for bringing the class action, and that was reduced to 12.5% ($187m) ahead of the ruling after objections from authors. Martínez-Olguín has ruled that this should be reduced to less than 7% of the settlement fund ($101m) and may be reduced further once accounting documents have been provided post-settlement distribution.
The three lead authors in the action saw their awards reduced from $50,000 to $15,000 after Martínez-Olguín labelled those requests “unreasonable”.
A handful of authors who decided to file separate lawsuits to seek higher damages tried to opt out of the settlement past the deadline and those requests were turned down.
Anthropic’s deputy general counsel Aparna Sridhar said the firm is happy with the landmark ruling that its AI training was fair use and that most rightsholders have already filed claims.
“We are pleased that more than 91% of authors and publishers covered by the settlement have claimed their share of the payment, and we’re looking forward to bringing this matter to a close,” Sridhar said.
The plaintiffs’ co-lead class counsel Rachel Geman. of Lieff Cabraser Heimann & Bernstein, LLP, said: “The settlement is historic. It reminds us, as if we needed reminding, of the paramount importance of books in reflecting and shaping the best of our language and thoughts, and the importance of those who write and publish them.
“Rightsholders came together to call foul. This is just the start.”
Lawsuits remain open against Google, Meta, Midjourney and ChatGPT maker OpenAI over whether it is legal to train AI models on copyrighted works.