Amazon Caught Destroying Rare Books to Train AI

Amazon has been found to be scanning and discarding a large collection of rare books as part of its AI training process. An investigation by 404 Media revealed that a bookseller suspected an order of 1,000 books from a marketplace with anonymous buyers was intended for an AI company. To confirm this, they placed an Apple AirTag inside one of the volumes to track its journey.

The shipment ended up at a large Amazon warehouse in Nevada, where workers admitted to stripping the spines off of books they receive to quickly scan their pages and destroy them. An Amazon employee on a worker forum confirmed their role, stating, "Some are assigned to cut books, and others go to receive where they get books and scan the barcodes."

Despite Amazon's claim that they purchase books through commercial channels to enhance their products and services, the investigation suggests the company is at the heart of a secretive AI industry practice. The practice has come to light through a lawsuit against Anthropic, which revealed similar methods used by the Claude chatbot maker.

A Nevada judge ruled that transforming physical texts into digital ones and disposing of the originals doesn't violate copyright law. Booksellers have long suspected that their products are being used and destroyed to train AI models, and the workers at the Amazon site's scanning of ISBNs supports this theory. Rare books, which have historical, intellectual, and sentimental value, are particularly targeted in this process, with the booksellers expressing concern that AI companies are disregarding the true value of these books.