Amazon has confirmed it buys books through commercial channels after an investigation linked a bulk order of rare books to a facility where physical copies are reportedly destroyed and scanned.
“Amazon purchases books through commercial channels to help develop and improve the products and services our customers use,” an Amazon spokesperson told City AM, offering few additional details.
The admission follows a 404 Media investigation that tracked roughly 1,000 books from an anonymous buyer to an Amazon facility in Las Vegas using an Apple AirTag hidden inside one of the volumes.
Workers at the site, known as VGT3, reportedly remove the spines of books before scanning their pages, rendering the physical copies completely unusable after the process is complete.
Employee posts cited by 404 Media also indicated that workers scan the books’ barcodes or ISBN numbers as part of their standard processing procedure at the facility.
Amazon did not specify which products the scanned material was being used to develop, nor did it explicitly confirm the books were being used to train its AI models.
Amazon develops its own family of Nova AI models and competes directly with OpenAI, Google, and Anthropic in the increasingly expensive race to build more powerful AI systems.
Printed and out-of-print books are particularly valuable to AI developers because much of their content has never been made available online, and older texts pre-date the explosion of AI-generated material on the internet.
That proliferation of AI-generated content has complicated efforts by developers to source reliable human-written training data, making physical books an increasingly attractive and strategically important resource.
Anthropic, the developer of the Claude AI assistant, had already been found buying millions of second-hand books under an internal programme known as Project Panama, cutting off spines and scanning the pages before recycling the originals.
A US federal judge subsequently ruled that Anthropic’s digitisation of legally purchased books amounted to fair use, distinguishing it from the company’s separate use of pirated copies.
Anthropic later agreed a $1.5bn settlement covering claims specifically relating to its use of pirated books, setting a notable legal precedent for the broader industry.
City AM reported last month that rare book dealers in London had received unusually large approaches from anonymous buyers seeking specialist and hard-to-find titles available only through specialist dealers.
One bookseller said they received two sizeable enquiries from email addresses carrying no names or company details, raising immediate suspicion about the true identity of the buyers.
The prospective buyers stopped responding after being asked to identify themselves and explain where the purchased books would be sent, suggesting deliberate concealment of their intentions.
Similar reports from booksellers across the market have raised concerns that AI developers or their suppliers are systematically targeting obscure and out-of-print titles that have never been digitised.
Publishers have begun turning that surging demand into a new revenue stream, with Amazon striking an AI licensing agreement with the New York Times last year, reportedly worth between $20m and $25m annually.