Amazon Destroys Rare Books to Feed AI Training Data Needs
Amazon, the company that began as an online bookseller, is purchasing rare books, cutting off their spines, and scanning them to train its AI models, according to a report from 404 Media. The outlet placed a tracking device inside a rare book and followed it to an Amazon facility in Las Vegas known as VGT3, which uses a dinosaur holding a book in its claws as its symbol. Amazon says it purchases books through commercial channels to improve the products and services its customers use.
Companies like Amazon need enormous amounts of text to train their large language models, and the internet has already been largely scraped for available material. Rare books, particularly those that are out of print or impossible to find online, represent a fresh and highly coveted source of training data for AI development.
Texts published before 2022 are especially valuable because there is no chance they were written by an LLM. When models train on AI-generated content, they risk model collapse, a phenomenon where output quality degrades after ingesting too much machine-generated text. Amazon's book-scanning operation highlights how far companies will go to obtain clean, human-written training data.