Amazon destroying rare books to scan them for AI training data

Quartz reports Amazon is buying rare books and destroying them to train its AI models. A 404 Media investigation tracked a rare book to an Amazon warehouse in Las Vegas where workers cut spines off books to scan them, destroying the physical coppies in the process. Rare and out-of-print books are valuable for AI training because they are not available anywhere else, providing insight into language and knowledge from a pre-digital era. 

The 404 Media investigation placed a tracking device inside a rare book and followed it to an Amazon warehouse were employees said their sole task is taking in bulk deliveries of books, slicing off the covers to run pages through scanners faster, leaving each volume unusable when the process is complete. The team’s logo features a dinosaur holding a book.

“We purchase books through commercial channels to improve the products and services customers use,” Amazon said in a statement, according to TechCrunch.

Amazon has been busy harvesting content from a variety of sources, including enrolling Twitch streamers in a program to let Amazon use channel content to train Gen AI models. Amazon owns Twitch, which automatically opted creators into the program. Creators could disable the setting.