Hidden Airtag reveals Amazon is trashing rare books to train AI

An investigator planted a hidden tracking device in a rare book purchased as part of a bulk order and traced it to a tech company's warehouse facility, confirming that the company was systematically destroying rare books after scanning them to train artificial intelligence models. The practice matters because it raises ethical concerns about the permanent loss of works that may have historical, intellectual, or sentimental value beyond their monetary worth, especially since booksellers and critics worry that unique texts will disappear while AI companies keep their training data secret. The tech company's statement indicated it purchases books to improve products and services but did not specifically address the destructive scanning practice, while some observers argue the books are obscure works that would otherwise sit unused, and others contend that even lesser-known texts warrant preservation. Competitors in the AI industry have publicly stated they are not training on rare or antique books, and booksellers suspect that AI firms are methodically targeting books by their identification numbers to maximize the volume of unique texts in their training datasets.