Amazon’s AI Data Hunt Is Turning Rare Books Into Scrap

Amazon’s AI Data Hunt Is Turning Rare Books Into Scrap
Amazon’s race for more AI training data is colliding with the physical fragility of the books that helped build its business. Rare and out-of-print volumes are reportedly arriving at a Las Vegas warehouse only to be dismantled, digitized and discarded.

The alarm grew after 404 Media placed a tracker inside a shipment of rare books and followed it to Amazon’s VGT3 facility in Las Vegas. The warehouse, according to the report, handles huge deliveries of printed material destined for high-speed scanning rather than resale or preservation.

Workers at the location said the process is blunt: “all they do is receive massive shipments of printed books which they then cut the bindings off in order to scan the books more quickly.” The result, they said, is that “the printed book is destroyed in the process.”

The books are valuable to AI companies precisely because many are scarce, offline and predate the wave of machine-generated writing. As the readily available internet becomes less useful as a clean source of human-authored text, older works can offer models material unlikely to have been produced by an AI system.

Amazon defended its acquisition practices in a statement to 404 Media, saying it “purchases books through commercial channels to improve the products and services customers use.” That frames the operation as a conventional supply-chain decision. But for booksellers and preservation-minded critics, the issue is not whether the books were bought legally; it is whether a one-time object should be treated as expendable once its words have been extracted.

The dispute exposes a sharp imbalance in the AI boom. A scanned text can be copied endlessly. A rare edition with its binding removed cannot be restored.

Continue reading https://foxvector.com/stories/01a01d00-3f76-1be1-7080-30b21ce4238d