Ilayda Cakirtekin
18 August 2026•Update: 18 August 2026
Amazon has been purchasing large quantities of books, scanning them to help train artificial intelligence models and destroying the physical copies in the process, according to an investigation published Monday by 404 Media.
The investigation found that books purchased in bulk through commercial marketplaces were sent to an Amazon warehouse in Las Vegas, Nevada, where workers reportedly removed their bindings and scanned their pages.
Workers cited in the report said the physical books were destroyed during the process.
404 Media tracked one book from a bulk order of about 1,000 titles after a bookseller agreed to place an Apple AirTag tracking device inside it.
The book was traced to Amazon’s LAS8 warehouse in Las Vegas, which also houses an operation known as VGT3.
According to workers cited by the outlet, VGT3 employees receive large shipments of printed books, scan their barcodes and pages, and remove the bindings to speed up the scanning process.
“Amazon purchases books through commercial channels to help develop and improve the products and services our customers use,” an Amazon spokesperson told 404 Media.
The report said the operation could be linked to growing demand for books as AI training data, noting that printed works contain large amounts of text that may not be readily available online.
Books published before the widespread adoption of generative AI are also considered particularly valuable as training material because they are less likely to contain AI-generated text. Researchers have warned that repeatedly training models on AI-generated material could contribute to declining model performance.
The investigation comes after booksellers reported a significant increase in bulk book purchases over the past year, with some suspecting that AI companies were acquiring large numbers of books for use as training data.
Sellers, however, were unable to identify the buyers because online marketplaces generally keep customer identities confidential.
