It was revealed that Amazon was purchasing and destroying rare books for artificial intelligence training, with an Airtag placed by a bookseller.
Suspicions that rare books were purchased in bulk and scanned and then destroyed in order to train artificial intelligence models have been confirmed by a new development. An Airtag placed inside a rare book by a bookseller proved that the book was headed to an Amazon artificial intelligence training facility in Las Vegas.
It was determined that in the facility in question, the books were separated from their bindings, their pages were scanned and then destroyed. In fact, it was noteworthy that the logo at the entrance of the facility included a T. rex figure eating a book.
In its statement on the subject, Amazon stated that it purchased the books through commercial channels and used them to develop its products. The company avoided directly addressing the issue of artificial intelligence training.
It is known that Amazon needs large amounts of original training data to remain competitive. In previous interviews with company employees, it was stated that the stocks of books to be scanned were exhausted and the facility was at risk of closure.
It is now certain that the facility continues its operations and the rare books that arrive are systematically destroyed. While employees state that this job is preferred due to its flexible hours, book lovers react to this destructive scanning method.
Research shows that artificial intelligence companies aim to methodically scan all printed books in the world using ISBN numbers. The fact that Amazon employees check barcodes or ISBNs before scanning books supports this theory.
Booksellers argue that this creates an ethical dilemma and that the valuation process is bypassed. AI companies are focusing on lesser-known or untranslated older books rather than high-value first editions.
Although these books have low monetary value, they can be of great historical and intellectual importance. Book sellers state that companies see the content of these works as just a pile of data.
On platforms such as Reddit, users discuss how problematic it is to destroy books sitting on dusty shelves. While some users argue that these books have not been read anyway, others are worried that historical information will remain locked in the hands of artificial intelligence companies.
Critics claim that AI models will only present censored or shared excerpts from these books. Some book sellers emphasize that works of no academic value can be sacrificed, but rare single examples should be preserved.
Companies' efforts to keep training data confidential make it difficult to provide transparency on this issue. Do you think it is acceptable to destroy rare books for artificial intelligence training?
