On July 20, 2026, US District Judge Araceli Martínez-Olguín granted final approval to a $1.5 billion copyright settlement from Anthropic, the AI company behind the Claude chatbot. The settlement resolves Bartz v. Anthropic PBC, a class-action lawsuit — a legal action brought on behalf of a large group of people with similar claims — that accused the company of copying hundreds of thousands of books without permission to train its artificial intelligence models.

The settlement covers roughly 482,000 copyrighted works. About 91% of the affected authors and publishers filed claims to receive a share of the payout. Bloomsbury Publishing, which had 14,087 titles listed in the case, said it expects to receive approximately $19 million (£14 million), pending deductions for legal fees and expenses.

The lawsuit began on August 19, 2024, when bestselling novelist Andrea Bartz, along with Charles Graeber and Kirk Wallace Johnson, filed a complaint (docket 4:24-cv-05417) in the US District Court for the Northern District of California. According to the Authors Guild, Anthropic downloaded approximately 7 million copies of books to train its AI models. A federal judge granted preliminary approval to the settlement on September 25, 2025. Judge Martínez-Olguín, who oversaw the final approval, stated the settlement provided "meaningful relief" to affected authors and publishers. The last known filing on the case docket is dated July 10, 2026.

Justin Nelson, the authors' lead lawyer, called the settlement "the largest known copyright recovery in history." The attorney fee award exceeds $100 million, though a federal judge previously reduced the plaintiffs' legal fee award, rejecting what was described as a "windfall" for the lawyers. The final settlement structure follows a May 14, 2026 court hearing where the judge considered the proposed agreement. Nearly 120,000 authors and other copyright holders sought a share of the $1.5 billion fund.

For Bloomsbury, the financial picture works out to a proposed baseline of about $3,000 per title. After a roughly 10% deduction for attorney fees and expenses, the publisher said it expects to receive the settlement cash in installments, potentially starting in the second half of this fiscal year. Bloomsbury confirmed the court's formal approval through a regulatory announcement via the London Stock Exchange's Regulatory News Service under ticker BMY. Anthropic's deputy general counsel, Aparna Sridhar, issued a written statement expressing satisfaction that more than 91% of authors and publishers covered by the settlement have claimed their share.

This settlement is the first major resolution out of dozens of AI copyright lawsuits filed in the US on behalf of authors and news outlets. The case docket also records filings from industry groups including the Textbook and Academic Authors Association, which filed a joint motion regarding settlement administration updates, and Anna Ganley of the Society of Authors, who filed a declaration in support of the preliminary approval motion.

The broader context here is that this settlement establishes a concrete financial baseline for unauthorized use of copyrighted works in AI training data. The per-title compensation figure — roughly $3,000 per book — provides a metric that will likely inform future litigation and legislative debates over how AI companies should license training data. Bloomsbury's parallel strategy is also worth noting: in 2025, the publisher announced an AI licensing deal allowing it to sell academic works to train generative AI programs, where authors could opt in and receive royalties. This dual track — seeking compensation for past unauthorized use while building future licensing frameworks — shows how publishers are navigating the intersection of intellectual property and machine learning, and it may well set a template that other rights holders follow.