Anthropic, an artificial intelligence company, has been accused of purchasing and destroying millions of physical books to train its AI models, including the chatbot Claude. This practice has raised significant concerns among authors, publishers, and the public regarding copyright infringement and the preservation of literary works.
In an effort to obtain high-quality training data, Anthropic reportedly acquired large quantities of used books from online retailers. The company then employed a process known as "destructive scanning," where books were disbound, their pages scanned into digital files, and the physical copies discarded. This method was intended to create a vast dataset for training AI models but resulted in the permanent loss of the original books.
The scale of this operation was substantial, with millions of books being processed in this manner. The destruction of these physical copies has sparked debates about the ethical implications of such practices, particularly concerning the preservation of rare or out-of-print titles. Critics argue that this approach prioritizes technological advancement over cultural and historical preservation.
In response to these actions, a group of authors and publishers filed a lawsuit against Anthropic, alleging copyright infringement. The case centered on the company's use of pirated books to train its AI models without proper authorization. In July 2026, a U.S. federal judge approved a $1.5 billion settlement in this landmark case, marking one of the largest copyright recoveries in history. The settlement provided compensation to affected authors and publishers, with approximately 91% of the claims already filed.
This case highlights the ongoing tension between technological innovation and intellectual property rights. As AI companies continue to develop advanced models, the methods used to source and process training data remain a contentious issue. The outcome of this lawsuit may set a precedent for future legal actions concerning the use of copyrighted materials in AI development.