A major dispute between artificial intelligence developer Anthropic and thousands of authors has ended in a $1.5 billion settlement. One of the largest recipients of the payouts will be the British publisher Bloomsbury, which publishes books by J.K. Rowling, Sarah J. Maas and other well-known writers. The agreement includes 14,087 works from the publisher, with compensation estimated at about $3,000 for each one.
The lawsuit was filed in 2024 and concerned the use of copyrighted books to train Anthropic’s language models. More than 91% of the 482,000 works covered by the settlement have already been claimed by rights holders for compensation.
This is the first major settlement in a series of lawsuits against companies developing generative artificial intelligence. It could become an reference point for other cases involving the legality of using books to train neural networks.
At the same time, Bloomsbury is independently developing licensing for AI content. The publisher already offers part of its scientific publications for model training, and authors who agree to take part in the program receive royalties. As a result, the market is gradually moving toward a model in which technology and copyright coexist on mutually beneficial terms.
