Amazon Is Destroying Rare Books to Train AI
Amazon started off selling books. Now they’re destroying them for AI training.
Let me break down what’s happening.
What’s Going On?
Amazon is buying rare books, cutting off their spines, and scanning them for AI training data. That’s according to a report from 404 Media, which placed a tracking device in a rare book that ultimately arrived at an Amazon facility in Las Vegas.
The facility, known as VGT3, identifies itself with a symbol of a dinosaur holding a book in its claws.
Amazon told 404 Media in a statement that it “purchases books through commercial channels to improve the products and services customers use.”
Why Rare Books?
Companies like Amazon need unfathomably large amounts of text to train their LLMs. They’ve already ingested everything they can from the internet.
Rare books offer a new source of coveted training data. Especially ones that are out of print or impossible to find on the internet.
And here’s the key reason. These texts are especially valuable because there’s no chance that anything published before 2022 was written by an LLM.
The Model Collapse Problem
When LLMs train on AI-generated text, they risk “model collapse.” That’s when the quality of an LLM’s outputs degrade after ingesting too much AI-generated text.
So companies are desperate for human-written text. The older and rarer, the better.
The Ironic Twist
Amazon started as an online bookstore. They built their business on books. Now they’re literally tearing them apart to build their AI.
It’s hard not to see the irony.
The Bottom Line
Amazon is destroying rare books to scan them for AI training data. They’re cutting off spines and digitizing texts that can’t be found on the internet. The company says they’re doing it to improve their products and services.
But the image of Amazon destroying rare books to train AI is not a good look.
