Under the settlement, eligible authors and publishers will receive $3,000 for each covered work, with payments spanning roughly 500,000 works. The court also approved more than $101 million in legal fees and rejected objections from some authors who argued the compensation was insufficient. According to Anthropic Deputy General Counsel Aparna Sridhar, more than 91% of eligible authors and publishers submitted claims.
Judge Araceli Martinez-Olguin issued the final approval after Judge William Alsup had granted preliminary approval in 2025 before retiring.
Earlier in the case, the court concluded that training an AI model on copyrighted books could qualify as fair use under copyright law. However, that finding did not extend to the way Anthropic assembled its book collection.
The court found that Anthropic obtained books through two different methods. Some titles were purchased and scanned legally, while others were downloaded from piracy websites, including Library Genesis and Pirate Library Mirror. According to the court, Anthropic stored more than seven million pirated books in a central repository, conduct that violated copyright law.
By agreeing to the settlement, Anthropic avoided a trial that could have exposed the company to additional damages related to its acquisition of copyrighted works.
“We reached this settlement in 2025, after the court’s landmark ruling that training AI on books is fair use under copyright law — which remains the law today,” Sridhar said in a statement.
Although the settlement brings this case to a close, it does not resolve the broader legal debate over whether AI companies can train models on copyrighted material. Because the case ended in a settlement after the district court ruling, the fair-use decision will not be reviewed by an appeals court and does not establish binding precedent for other courts.
Some authors and publishers chose not to participate in the settlement and continue to pursue separate lawsuits against Anthropic.
The decision also comes as similar copyright disputes continue across the AI industry. Companies including Google, Meta, Midjourney, and OpenAI face lawsuits challenging the use of copyrighted works for AI training. More recently, publishers and authors including Hachette, Cengage, Elsevier, Scott Turow, and S.C.R.I.B.E. filed a class action lawsuit against Google, alleging the company used copyrighted books to train its Gemini AI platform.
Separately, Anthropic is pursuing a trademark infringement lawsuit against AI security company Abnormal AI, alleging that one of the startup’s rebranded logos is too similar to its own.
This analysis is based on reporting from The American Bazaar.
Image courtesy of WinBuzzer.
This article was generated with AI assistance and reviewed for accuracy and quality.
