A federal judge has approved a $1.5 billion settlement between Anthropic and a group of authors who claimed the company trained its AI models on their copyrighted books without permission. Judge Araceli Martínez-Olguín signed the order on Monday after reviewing the final agreement reached with the plaintiffs. The ruling states the deal offers meaningful relief to the writers, providing approximately $3,000 for each book the company allegedly used in its training data. This agreement marks the largest known copyright recovery in history according to the law firm representing the authors. The lawsuit was originally filed by Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson who argued that Anthropic’s systems scraped their work to improve its large language models. The settlement resolves the legal dispute without requiring a full trial on the technical claims regarding data scraping methods. While the total sum is substantial, individual payouts depend on the number of copies of each specific book used by the model. This decision sets a significant financial precedent for future artificial intelligence copyright disputes in the United States.
- Total payout is capped at $1.5 billion for all participating authors combined
- Individual compensation is calculated at roughly $3,000 per allegedly pirated book
- The settlement avoids a lengthy trial on the technical details of data collection
