Anthropic’s $1.5B copyright settlement approved; only 350 authors opted out

Key Points:

• A judge approved a $1.5 billion settlement between AI company Anthropic and authors, marking the largest copyright class-action and settlement in history, following a ruling that Anthropic’s AI training on books was fair use but piracy was not.
• Despite objections from some authors over low payouts (around $3,000 per work) and high lawyers’ fees, the court ruled the settlement was fair, with about 91% of affected authors and publishers filing claims and only a small fraction opting out.
• The judge reduced lawyers’ fees from the requested $300 million (20%) to about $101 million (less than 7%) and cut the service awards for lead plaintiffs from $50,000 to $15,000, citing concerns over inflated fee projections and lack of retaliation risk.
• Anthropic successfully blocked most late opt-out requests, including from notable author Donald Passman, with the court emphasizing the importance of finality and noting that timely notice was sent to most class members.
• The settlement includes non-monetary benefits such as Anthropic’s obligation to destroy the authors’ works used in training and provisions for future litigation if misuse occurs, providing additional protections beyond the financial compensation.