US judge gives preliminary approval to Anthropic $1.5bn proposed AI training copyright settlement after getting assurances all claimants will get fair pay-out

Photo by Bermix Studio on Unsplash

The US Federal judge hearing the court case over AI company Anthropic’s unauthorised use of copyright material in training its AI model, Claude, has now given preliminary approval for a proposed $1.5bn settlement – the biggest award ever in copyright law history – to go ahead.

According to Reuters, Maria Pallante, president of the publishing-industry trade group the Association of American Publishers, said the settlement was "a major step in the right direction in holding AI developers accountable for reckless and unabashed infringement."

Judge William Alsup had asked for reassurances that all legitimate claimants would get fair treatment.

In June 2025, he delivered a mixed verdict in which he ruled that using copyright material to train AI was not in itself illegal; however, Anthropic broke the law by accessing copyright material from pirate websites, he said, and should stand trial for that.

Under the terms of the proposed settlement, Anthropic would pay $3,000 for each piece of copyrighted work that was pirated and avoid any trial. Anthropic deputy general counsel Aparna Sridhar said that the decision will let the company "focus on developing safe AI systems that help people and organizations extend their capabilities, advance scientific discovery, and solve complex problems."

The lawsuit – Bartz v Anthropic – was originally filed in August 2024 on behalf of three authors whose work had been downloaded from pirate sites. Thriller writer Andrea Bartz is one of three authors who were the original plaintiffs; the others are Charles Graeber, who wrote true crime book The Good Nurse author and Kirk Johnson , who wrote true crime book The Feather Thief. However, the case is now a class action on behalf of potentially many thousands of authors and publishers.

Andrea Bartz, Charles Graeber and Kirk Wallace Johnson said in a statement that Alsup's decision "brings us one step closer to real accountability for Anthropic and puts all AI companies on notice they can’t shortcut the law or override creators’ rights.”

Anthropic, an AI developer backed by big tech firms including Amazon and Alphabet, parent company of Google, was claimed to have a library of nearly 500,000 pirated books, according to a lawyer for the plaintiffs quoted in The Independent.

Judge Alsup said during a hearing on Thursday last week that the proposed class-action settlement was fair. However, he will only decide whether to give the settlement his final approval after notifying affected authors and giving them a chance to file claims.

The issue of how AI developers have used copyrighted material in training their AI models has sparked major legal battles, not just over written material but also for film, photography and visual arts, music and even personality rights. How the Anthropic case plays out is being seen as a watershed for AI development.