Judge approves $1.5B Anthropic settlement for pirated books used to train Claude (apnews.com)

A federal judge has approved a landmark $1.5 billion settlement for Anthropic, stemming from the use of pirated books to train its Claude chatbot. Each author affected will receive about $3,000 for their work, and a significant 91% of the over 482,000 books involved have been claimed. This ruling is notable as it marks the largest copyright recovery in history and offers "meaningful relief" to authors, according to District Judge Araceli Martínez-Olguín. Despite a mixed initial ruling from Judge William Alsup, who determined training AI on copyrighted material can be fair use, he noted that Anthropic improperly sourced millions of books from pirate sites. This case, initiated by bestselling author Andrea Bartz and others in 2024, sets a crucial precedent amid ongoing AI copyright lawsuits, underscoring the legal complexities surrounding the use of copyrighted materials in machine learning. Anthropic's response indicates a potential shift in the understanding of copyright in relation to AI development, making the case significant for the AI/ML community and future litigation.