TLDRocket
Sign in

Inside the Warehouse Where Amazon Scans and Destroys Books for AI Training

404 Media Emanuel Maiberg

An Amazon worker says books at a Las Vegas warehouse were scanned for AI training, then cut apart and tossed. The weird part: some looked brand new, and even rare or library books were in the mix.

Based on reporting by 404 Media, Emanuel Maiberg — read the original for the full story.

Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error

At Amazon’s VGT3 warehouse in Las Vegas, the job wasn’t just scanning books. According to one worker who spoke anonymously to 404 Media, books were unpacked, sorted, fed into scanners, and then physically destroyed by having their spines cut off. The pages were then dumped loose into large shuttles, Amazon’s term for big cardboard bins, so the books couldn’t be put back together.

The worker said the operation sat in the back of the warehouse, past the scanners, and that even some employees seemed unsure what the place was doing. That uncertainty didn’t last long. After seeing the cutting and scanning process, the worker said it became clear the books were being handled for AI training, not for Kindle or any kind of normal book service.

The books themselves sounded like a strange cross-section of the world. Some were brand new. Some looked like they had come from libraries. The worker mentioned boxes from London, books from the University of London, German and Russian titles, and whole pallets of Japanese books still sealed. There were also government-looking documents, including stapled papers presented to Parliament on behalf of Her Royal Majesty the Queen.

The workflow was blunt. Books came off trucks, were staged, sorted into totes, and moved to the people who removed duplicates and sent the rest on for scanning. Then came the spine-cutting machines, one worker at a station, a cover over the blade, and a button press that split the book open. The pages were scanned fast, with the screens showing the books flipping by, and then the loose sheets went into shuttles with no way back.

The employee said the process changed every day and didn’t look well managed. But the bigger issue was simpler than warehouse chaos. These were books, not just data objects. Once Amazon turns them into training material, the physical copies are gone, which means less access for everyone else.

My take — AI-written commentary, not fact-checked reporting

This is the sort of AI story that should make publishers and librarians nervous, because the industry keeps acting like every book is just fuel for the machine. Scanning rare and library books, then shredding the originals, has all the grace of a paper shredder in a museum. The AI boom has a bad habit of calling that progress.

Read more about this at: 404 Media

Related stories

The daily briefing

Every AI story that matters, in your inbox by 8am.

TLDRocket reads all relevant sources, removes duplicate coverage, and summarises the day in two minutes. Follow companies and topics for alerts, or get the briefing in Slack. Free, no spam, unsubscribe anytime.