Euronews report: Anthropic destroyed millions of books in $1.5B deal

Euronews has released a detailed reconstruction alleging that Anthropic, the AI company behind Claude, secretly scanned and destroyed millions of physical books as part of its model training pipeline. The report claims these books were obtained and processed without proper authorization, raising fresh concerns about how leading AI labs acquire training data. The alleged mass digitization-and-destruction operation was reportedly designed to avoid detection and bypass copyright protections.

This bombshell comes as Anthropic is finalizing a $1.5 billion copyright settlement with publishers, a figure that underscores the scale of the alleged infringement. While Anthropic has not publicly confirmed the details, the Euronews reconstruction paints a picture of systematic, covert data collection that could have lasting legal and reputational consequences. The case highlights the growing tension between AI innovation and intellectual property rights, and it may set a precedent for how other AI companies handle copyrighted materials.