Judge Approves Anthropic's $1.5 Billion Copyright Deal, Cuts Lawyers' Fees to $101 Million
Updated
Updated · Ars Technica · Jul 21
Judge Approves Anthropic's $1.5 Billion Copyright Deal, Cuts Lawyers' Fees to $101 Million
3 articles · Updated · Ars Technica · Jul 21
Summary
$1.5 billion in payments to authors and publishers won final court approval Monday, ending the largest certified copyright class action against Anthropic.
Judge Araceli Martínez-Olguín rejected objections over low payouts and notice, citing 95% class notification, 91% claims filed, only 350 opt-outs, and 54 objections or late opt-out requests.
The settlement stems from a ruling that Anthropic's AI training on books was fair use but its piracy of works likely was not; the deal pays about $3,000 per work—four times minimum statutory damages.
Martínez-Olguín also slashed attorneys' fees from a revised 12.5% request of about $187 million to under 7% of the fund, roughly $101 million, and reduced awards for the three class representatives.
Anthropic paid $1.5B, but why do many authors see the landmark AI settlement as a loss for creators?
With over 100 lawsuits pending, will AI's future be built on licensing deals or courtroom battles?
$1.5 Billion Anthropic Copyright Settlement Sets New Precedent for AI Data Sourcing and Creator Rights
Overview
In August 2025, Anthropic agreed to a landmark $1.5 billion settlement after facing a major class-action lawsuit that began in August 2024. The lawsuit accused Anthropic of using pirated copies of about 500,000 copyrighted books from illicit online sources to train its AI models without permission. As a result, both authors and publishers became eligible for payouts from the settlement fund. To ensure fair distribution, an Author-Publisher Working Group was proposed to manage how the funds would be allocated, marking a significant moment for copyright and AI in the publishing industry.