Inside Amazon's Book Scanning Warehouse: A Glimpse into AI's Dark Side

An anonymous employee from Amazon's VGT3 warehouse in Las Vegas has lifted the curtain on a practice that many might find unsettling: the scanning and destruction of thousands of books for AI training. This revelation follows a recent investigation that tracked a shipment of rare books suspected to be headed for the tech giant's facilities.

The VGT3 warehouse, part of a larger Amazon complex, operates alongside a print-on-demand facility. According to the employee, the process begins with the arrival of boxes filled with books, some brand new, others used. The books range from an eclectic mix of languages to unusual items like government documents presented to the UK Parliament.

Upon entering the warehouse, the employee recalls a sense of confusion. "I saw all the scanners and wondered what was going on. I asked someone about it, and they seemed uncertain too. It became clear to me that this was not just a regular operation," they shared.

The process involves unboxing the books, cutting off their spines, and scanning the loose pages—an operation that ultimately reduces these literary works to mere data. The employee noted, "After scanning, they throw all the loose leaf pages into big shuttles. There’s no way to put the book back together." This method of destruction is not just a matter of efficiency; it signifies a troubling disregard for the original works.

The employee described the chaotic environment at VGT3. "The process changed every day, and it seemed poorly managed. We’d stack the books into totes for scanning, and duplicates were tossed into shuttles, often in a haphazard manner. They said they were returning them to vendors, but I’m skeptical about the actual process."

In the cutting area, specialized machines are used to sever the spines from the books. The employee explained, "It’s a safe device. You slide the book in, remove your fingers, and press a button. It slices the book, and then you remove the pieces. After that, the pages are tossed into shuttles."

Curiously, the employee reflected on the purpose of this operation. Initially, they thought the books were intended for Kindle devices, a notion they quickly dismissed upon learning they were for AI training. "I don’t like it. There’s so much knowledge in these books, and many might be rare. It’s disheartening to think they can’t be reused," they lamented. The practice raises questions about the ethical implications of data collection in the tech industry and the potential loss of literary heritage for the sake of artificial intelligence.

Source: www.404media.co