2 weeks ago
Amazon Reportedly Scanning Books for AI Training Data
Amazon is buying lots of books and scanning them to help teach its AI programs.
The books contain information that is hard to find online, and their neat format makes them easy for computers to read.
Some of the books get destroyed after they are scanned.
A news website called 404 Media put a tiny tracker inside a rare book and followed it to an Amazon warehouse in Las Vegas.
The workers there said they did not know much about what the books were for.
This is similar to something another company called Anthropic did, and a judge said it was okay, but many authors are upset about it.
Amazon is also changing its AI plans.
It is stopping some of its own AI models and reorganizing its teams.
Some Amazon workers lost their jobs because of these changes.
A new leader, Pieter Abbeel, is now in charge of Amazon's most important AI research group.
Amazon is reportedly buying large quantities of books and scanning them for AI training data, according to a 404 Media investigation.
Investigators placed a tracking device inside a rare book that eventually ended up at an Amazon warehouse in Las Vegas.
Workers at the facility said they had little information about the operation and processed massive shipments of printed books to enable faster scanning.
The practice drew criticism after a lawsuit against Anthropic revealed Project Panama, in which books had their spines cut off before being scanned; a judge ruled the practice fair use.
Amazon is restructuring its AI strategy, winding down several Nova models including Premier, Omni and Reel, and making the Frontier Model Research group, led by Pieter Abbeel, a key priority.
- Who
- Amazon, its warehouse workers in Las Vegas, and AI researchers
- What
- Buying and scanning large quantities of books for AI training data, with books sometimes destroyed in the process
- Where
- An Amazon warehouse in Las Vegas
- When
- Recent investigation; Rohit Prasad left Amazon in December 2025
- Why
- To gather training data for AI models, since books contain detailed information that is often difficult to find online
Critics
AI Companies and Courts
Scanning copyrighted books for AI training
Critics
Authors, activists and e-reader communities criticize the practice, arguing books are being used without permission and ultimately destroyed.
AI Companies and Courts
AI companies defend the practice, and a judge ruled that scanning books for AI training constitutes fair use.
Amazon's AI restructuring
Critics
The restructuring has led to layoffs and the winding down of several in-house models, raising concerns about disruption.
AI Companies and Courts
Amazon is consolidating its AI efforts to focus resources on competing at the AI frontier.
Key facts
- Company
- Amazon
- Investigation
- 404 Media
- Warehouse location
- Las Vegas
- Related program
- Project Panama (Anthropic)
- Legal ruling
- Fair use
- Wound-down models
- Nova Premier, Nova Omni, Reel
- Leadership change
- Rohit Prasad left in December 2025
- Frontier Model Research lead
- Pieter Abbeel
Sources
Related news
Anthropic’s Destructive Book Scanning Raises Copyright and Preservation Fears
AI firms buy second-hand books for destructive scanning surge
Groups Urge FTC Probe Into AI Firms Destroying Books
Anthropic destroying books to make AI data: an existential challenge
Great Book Grab: Who Owns Knowledge in the AI Age?
