An AirTag hidden in a rare book shipment led investigators to an Amazon facility in Las Vegas where workers cut spines and scan pages to build AI training data, according to Cybernews and 404 Media.
Amazon told reporters it acquires books to “help develop and improve the products and services our customers use.” It did not detail the scale of the program or say why scanning requires destroying the books.
Amazon is not alone. Anthropic’s internal documents describe a program called Project Panama, an effort to “destructively scan all the world’s books,” with one vendor proposal seeking between 500,000 and two million books over six months, per Betakit’s reporting on the 404 Media investigation. That single request would exceed the roughly 1.2 million volumes held by Montréal’s Grande Bibliothèque.
A Dutch Bookseller’s 3,000-Copy Order
Fortune reported that a Dutch secondhand bookseller received a request for 3,000 copies of a single title and initially suspected phishing. NPR frames the pattern as part of a broader scramble for fresh text data as large language models exhaust available digital sources.
Publishers, authors and preservationists have objected to the destruction. One bookseller told the BBC that the world
Sources
- Betakit: contributed reporting
- TechCrunch: contributed reporting
- NPR: contributed reporting
- Fortune: contributed reporting
- Cybernews: contributed reporting
Reporting compiled and contextualised by TechScoop Canada. Figures reconciled across the sources above.
Originally reported by Betakit.