A US federal judge has approved a landmark US$1.5 billion copyright settlement requiring AI company Anthropic to compensate thousands of authors and publishers whose books were used without permission to train its Claude AI model.
Earlier this week, US District Judge Araceli Martínez-Olguín granted final approval to the agreement, bringing to a close a two-year long class action lawsuit brought by a group of authors who alleged that Anthropic had downloaded hundreds of thousands of pirated books from illegal online repositories, including LibGen and PiLiMi, to train its AI systems.
The settlement covers more than 491,000 works, with eligible authors and publishers expected to receive approximately US$3,000 per work. Anthropic has also agreed to destroy all copies of books obtained from known pirate libraries, marking an important step towards greater accountability in the development of generative AI.
Although the settlement relates specifically to the unauthorised use of pirated books, it represents an important milestone for authors’ rights and copyright enforcement in the age of artificial intelligence.
The International Authors Forum welcomes this important development and will continue to advocate for an AI ecosystem founded on transparency, consent, and fair licensing practices. Authors must have the right to know when their works are being used, to authorise that use, and to receive fair remuneration for their creative contribution.