Amazon is acquiring old, obscure, and rare books and sending them to a Las Vegas facility for destructive scanning as part of AI-related work, according to an investigation by 404 Media reported by medianama.com. The process involves removing book spines and separating pages for faster scanning, after which the physical copies are discarded. This practice raises concerns about the loss of irreplaceable books, especially those out of print or with unique physical features.
404 Media tracked a rare book by placing a tracking device inside it and followed its shipment to Amazon's VGT3 facility in Las Vegas. Amazon confirmed it purchases books through commercial channels to improve its products and services but did not disclose the volume of purchases or whether rare titles are preserved. Booksellers and preservationists have voiced worries about the destruction of physical copies, which cannot be fully replicated by digital files due to handwritten notes, bindings, or other unique attributes.
The issue extends beyond Amazon, with European booksellers reporting unusual bulk orders of thousands of unrelated and obscure titles, suspected to be from AI companies or intermediaries. An Irish bookseller, Kennys, flagged an order for 5,000 obscure titles, highlighting a trend that diverges from typical library or collector purchases. The destruction of rare books for AI training data underscores tensions between digitization efforts and cultural preservation.
The investigation by 404 Media highlights the scale and impact of this practice, with Amazon's VGT3 facility in Las Vegas identified as a key site for scanning and discarding physical books. The concerns raised by booksellers and preservationists emphasize the need for transparency and safeguards in AI training data acquisition.