08/23/2026
This is only one story about this. It has been in the news on several stations. Be aware of what we are working against or the day will come that you won’t have old books to be reading.
A single Apple AirTag may have helped reveal how physical books are being turned into data for AI training.
An investigation by 404 Media, conducted with an independent bookseller, tracked a shipment from a large anonymous order of roughly 1,000 books. An AirTag hidden inside one of the books traveled across several states before arriving at a high-security Amazon facility in Las Vegas, Nevada.
According to the investigation, the facility uses a specialized operation to process books at scale.
The books are reportedly stripped of their bindings, their pages are scanned and digitized, and the physical copies are then destroyed.
The reason this matters is the growing demand for large collections of high-quality text to train artificial intelligence systems.
Books published before the widespread arrival of generative AI can provide models with human-written material that was created independently of today’s AI systems. Rare and obscure books can be particularly valuable because much of their content may not already exist in easily accessible digital databases.
For independent booksellers, however, the process raises uncomfortable questions.
A physical book can be legally purchased as an ordinary commercial product. But what happens after it is bought, dismantled and converted into a digital dataset for AI development is a much more complicated issue.
It touches on copyright, licensing, privacy, ownership and the broader question of whether purchasing a physical copy gives a company the right to transform its contents into training material.
Amazon has defended its book-purchasing activities, saying the company acquires books through commercial channels to improve its products and services.
The investigation does not establish that every book purchased by Amazon is being used for AI training, nor does it settle the legal questions surrounding the practice.
But the image is difficult to ignore:
A book arrives at a warehouse.
Its binding is cut away.
Its pages become data.
And the physical book disappears.
As AI companies race to acquire more information, the newest frontier of data collection may not be hiding on the internet.
It may be sitting quietly on a bookshelf.
