Amazon Destructively Scans Rare Books for AI Training
Amazon operates a Las Vegas facility that destroys rare and out-of-print books to scan them for AI training data, according to a 404 Media investigation.
An investigation by 404 Media revealed that Amazon.com, Inc. operates a facility in Las Vegas where rare and out-of-print books are destructively scanned to train its frontier artificial intelligence models. Using an Apple AirTag planted in a bulk order of 1,000 volumes sourced via the Biblio marketplace, reporters tracked the shipment to a unit coded VGT3 within the LAS8 warehouse.
Internal reports and worker forum posts describe a process where staff slice the spines and bindings off books to feed pages through high-speed scanners before discarding the physical copies. The facility's team reportedly uses a logo featuring a Tyrannosaurus rex devouring a book. Amazon stated that it purchases books through commercial channels to improve its products and services, but it has not specifically addressed the destructive nature of the scanning.
AI companies target pre-2022 printed texts to avoid model collapse, a phenomenon where AI performance degrades when trained on AI-generated web content. This practice mirrors Project Panama, a similar initiative by Anthropic PBC. While a 2025 U.S. District Court ruling in Bartz v. Anthropic suggested such scanning constitutes transformative fair use, the legality of the practice remains contentious in Germany and the United Kingdom.
Keep reading in the app
The full story and every source, free in the app.
Coming soonGoogle Play