U.S. District Judge William Alsup has refused to give preliminary approval to a record‑breaking $1.5 billion settlement that Anthropic reached with a class of authors alleging the AI company used pirated works to train its large language models. Alsup said the deal was “nowhere close to complete” and raised concerns about inadequate notice to class members, a vague list of works, and the lack of a clear opt‑in process. He gave the parties a deadline to submit a finalized list of works by September 15 and required court‑approved notice and claim forms by early October before any settlement can move forward.
Read more →