Big tech faces historic copyright reckoning in landmark AI case

Judge Araceli Martínez-Olguín has granted final approval to the $1.5 billion settlement in Bartz v. Anthropic, concluding a major copyright case involving AI training practices. The ruling found the settlement "fair, reasonable, and adequate," with payments beginning within approximately one month.

The settlement covers nearly 500,000 works, with authors receiving up to $3,000 per work. The claims rate reached 91.3%, with only 53 objections filed.

The case arose from Anthropic's use of copyrighted books to train its Claude AI system. Judge William Alsup previously found the training itself constituted fair use but determined that Anthropic's retention of unauthorized downloads in a permanent library likely infringed copyright, prompting settlement negotiations.

The outcome establishes important parameters for how publishers and technology companies navigate copyright law and AI development, providing a framework for similar pending cases.

Payments begin one business day after the appeal deadline expires—approximately 30 days. The settlement distinguishes between authors and publishers, with larger publishers receiving substantially higher payments reflecting their catalog scope.

Judge Martínez-Olguín reduced attorney fees to approximately $101.5 million—a significant cut from the plaintiffs' lawyers, who initially sought $300 million before trimming their request to $187.5 million. While 25% of a common fund is typically reasonable for class actions, the court held this standard doesn't necessarily apply to "mega-funds."

"This amounts to nearly 6.8% of the settlement fund, which is a reasonable percentage given a fund of this size," the judge ruled.

Based on 34,381.6 credited hours, the award equals approximately $2,953 per hour. Service awards for named plaintiffs were reduced from $50,000 to $15,000 each.

Maria Pallante, CEO of the Association of American Publishers, stated: "We applaud the court's final approval of this settlement, which represents an important victory in the larger battle to hold big tech accountable for its unscrupulous appropriation of intellectual and creative properties."

The settlement may influence ongoing litigation, providing non-class members with a validated legal theory and damages benchmark.

The settlement establishes precedents that will shape future disputes between rights holders and AI companies. The court's distinction between training use (fair use) and permanent retention of unauthorized copies (likely infringing) provides a framework acknowledging both innovation and copyright protection.

The high claims rate suggests broad industry support for negotiated resolutions. Non-class members retain independent legal rights, and with multiple cases proceeding separately, the publishing industry's legal engagement with AI training practices continues.