Bloomsbury, the UK publisher known for the Harry Potter series, is set to receive approximately £30 million as part of a landmark $1.5 billion (£1.1 billion) copyright settlement with US-based AI company Anthropic. The settlement resolves claims brought by thousands of authors who alleged that Anthropic used their copyrighted works without permission to train its Claude chatbot AI models.

The legal action, initiated in 2024 by authors including Andrea Bartz and others, covered a total of 482,000 copyrighted works, of which about 91% have been claimed by rightsholders. Bloomsbury reported that it had 14,087 titles included in the settlement, representing popular authors such as Sarah J Maas, Susanna Clarke, and J.K. Rowling. Compensation is estimated at roughly $3,000 (£2,200) per title, before deductions for legal fees and other expenses, with the payout split evenly between the authors and the publisher.

The case highlights ongoing tensions in the industry over AI training data, particularly the use of extensive copyrighted material sourced from the open web without explicit licensing. While US AI firms, including Anthropic, have argued that their use of such content falls under the “fair use” doctrine—which permits limited use of copyrighted material without permission—many authors and publishers have challenged this stance, demanding appropriate licenses and compensation.

A US district court judge, Araceli Martínez-Olguín, described the settlement as providing “meaningful relief” to those affected. Justin Nelson, the lead attorney representing the authors, called it the “largest known copyright recovery in history.” The settlement represents one of the first significant resolutions among numerous ongoing copyright lawsuits targeting AI companies for unauthorized use of literary and journalistic works.

Anthropic’s deputy general counsel, Aparna Sridhar, expressed satisfaction with the resolution, noting that over 91% of eligible claimants had opted to receive compensation and that the company looked forward to closing the matter.

Bloomsbury itself has pursued a proactive approach by establishing a licensing framework last year, allowing it to sell academic and literary content for use in training generative AI models. Authors under Bloomsbury’s umbrella have been given the option to participate in this scheme and receive ongoing royalties in exchange for permitting their works to be used for AI training.

The settlement is viewed as a significant development in the evolving debate over copyright and artificial intelligence, balancing the rights of content creators with the technological advancements of AI systems. The outcome may influence future licensing practices and litigation related to AI-generated content and intellectual property rights.