DMR News

Advancing Digital Conversations

Amazon Is Buying and Destroying Rare Books for AI Training, Report Says

ByJolyen

Aug 18, 2026

Amazon Is Buying and Destroying Rare Books for AI Training, Report Says

Amazon is buying large quantities of rare books, scanning their contents for artificial intelligence training data and destroying the physical copies in the process, according to an investigation by 404 Media. The publication traced one shipment containing a tracking device to an Amazon facility in Las Vegas used for scanning books.

The 404 Media investigation identified the facility as VGT3, which uses an image of a dinosaur holding a book as its symbol. Amazon told the publication that it purchases books through commercial channels to improve products and services used by customers.

The reported process involves removing the spines of books so their pages can be scanned in bulk. Many of the books being acquired are rare, out of print or otherwise difficult to find in digital form, giving Amazon access to text that may not already be readily available online.

Rare Books Offer Another Source of Training Text

Large language models require substantial amounts of text during training, including material from books, articles and other written sources. Amazon itself has previously described generative AI systems as models trained on billions of data points from sources that include books and articles.

Rare and older books can provide material that is absent from online datasets. Books published before the widespread adoption of generative AI also offer human-written text that predates the surge in AI-generated material.

That distinction matters because researchers have studied a problem known as model collapse, in which repeatedly training models on generated data can reduce the quality and diversity of subsequent outputs. The concern increases as more AI-generated text appears on the internet and becomes part of datasets collected for future models.

Amazon Has Not Detailed the Models Using the Scans

Amazon has not publicly identified which specific AI models are being trained with text gathered through the book-scanning operation, according to 404 Media. Its public AI resources cover a range of generative AI products and research, but do not describe the VGT3 book-processing program.

The Las Vegas operation also differs from simply collecting digital copies of books because the company is purchasing physical volumes and converting them into machine-readable data itself. 404 Media said its investigation revealed an Amazon book-buying operation that had not previously been reported.


Featured image credits: Wikimedia Commons

For more stories like it, click the +Follow button at the top of this page to follow us.

Jolyen

As a news editor, I bring stories to life through clear, impactful, and authentic writing. I believe every brand has something worth sharing. My job is to make sure it’s heard. With an eye for detail and a heart for storytelling, I shape messages that truly connect.

Leave a Reply

Your email address will not be published. Required fields are marked *