Authors celebrate “historic” settlement coming soon in Anthropic class action (arstechnica.com)

Authors and Anthropic are on the verge of finalizing a groundbreaking settlement in a class-action lawsuit alleging unauthorized use of copyrighted books to train Anthropic’s AI models. U.S. District Judge William Alsup confirmed that both parties have agreed "in principle" to the settlement, with a motion for preliminary approval expected by early September. The lawsuit, initially filed by three authors, was notably expanded to potentially include up to seven million claimants due to the widespread use of copyrighted texts in Anthropic’s dataset—a move that AI advocates warned could financially devastate the AI industry. This settlement marks a critical moment for the AI/ML community as it underscores the legal and ethical challenges around training data sources for large language models. While details remain confidential, the binding nature of the terms indicates a precedent-setting resolution that might influence how companies source and clear data going forward. The case highlights growing tensions between content creators and AI developers, emphasizing the need for clearer frameworks around data rights and compensation in AI training. The outcome could shape industry practices, potentially affecting how AI startups approach scalability and compliance amid escalating copyright scrutiny.