For months, booksellers suspected that technology companies were buying large lots of rare books and destroying them after scanning the contents for artificial intelligence training. Concrete proof remained elusive until a hidden tracking device revealed that at least one major player was behind the practice: Amazon.

What You Need to Know

Investigators working with 404 Media used a Hidden Airtag to track a bulk shipment of rare books to an Amazon AI training facility in Las Vegas. Workers at the facility were seen tearing books from their spines and scanning pages for data. Amazon declined to comment on the findings and did not confirm that the materials were used for AI development. The discovery adds to growing scrutiny over how tech giants source training data for frontier models.

The Investigation

On Monday, 404 Media revealed that a bookseller agreed to plant a tracking device inside a rare book as part of a bulk order. That Airtag was then tracked to a warehouse in Las Vegas operated by Amazon AI. Investigators documented that the facility housed a team focused on destroying books and scanning their pages. A logo on the door depicted a T. rex devouring a book.

Key findings from the investigation include:

  • Hidden Airtag: A bookseller placed the device in a rare book as part of a bulk order.
  • That Airtag: It was tracked directly to an Amazon AI training facility in Las Vegas.
  • Destructive process: Workers tore books from their spines and scanned each page for data collection.

Amazon's Response

Amazon declined to comment on the 404 Media report. The company provided only a brief statement that did not mention AI training or the destruction of books. The lack of transparency has drawn criticism from the publishing community and advocates for archival preservation.

The discovery comes amid a broader race among tech giants to secure high-quality training data for large language models. Rare books are valued for their unique content and historical significance. Destroying them for data extraction has sparked outrage among booksellers and researchers.

Why This Matters

This event exposes a troubling practice at the intersection of AI development and cultural heritage. If Amazon AI and other firms routinely destroy rare books to train models, the loss of physical artifacts becomes a permanent cost of technological progress. Booksellers face an ethical dilemma: sell to AI companies and risk losing irreplaceable works, or refuse and lose revenue. Regulators and watchdog groups may push for stricter oversight of how training data is sourced, especially when it involves irreplaceable items.

The investigation also demonstrates the power of grassroots tracking methods. A simple Hidden Airtag provided evidence that corporate denials and speculation could not. It raises questions about how many other rare collections have been quietly dismantled for AI training without public knowledge.