Judge: Anthropic’s $1.5B settlement is being shoved “down the throat of authors”

During a court hearing on Monday, U.S. District Judge William Alsup expressed serious concerns regarding a proposed $1.5 billion settlement related to Anthropic's alleged unauthorized use of books for AI training. This settlement comes in light of a case where the potential damages could exceed a staggering $1 trillion, as Alsup had approved a class that includes up to 7 million authors whose works were reportedly downloaded without permission. Critics argue that the settlement allows Anthropic to escape with a relatively small penalty, covering less than 500,000 works while paying only a fraction of its current valuation of approximately $183 billion. Notably, the terms of the settlement do not require Anthropic to admit any wrongdoing, even as the company continues to secure substantial funding—most recently raising $13 billion, which is nearly ten times the amount of the proposed settlement. Judge Alsup conveyed his discontent with how quickly the lawyers rushed the agreement, warning that it risks being forced "down the throat of authors." He pointed out the myriad unresolved issues within the proposed settlement, including the Works List, Class List, Claim Form, and the processes for notification and dispute resolution. By denying preliminary approval, Alsup indicated that the current state of the agreement is far from satisfactory and necessitates a significant overhaul. A key change Alsup is advocating for is the inclusion of a requirement for all copyright holders to opt in to the settlement. He emphasized that if even one rights holder chooses to opt out, their work would then be excluded from the settlement. Additionally, he insisted that any disputes regarding ownership or claims should be resolved in state court, highlighting the need for a clearer framework before any approval can be granted.