AI Firms Are Quietly Buying and Destroying Millions of Printed Books to Train Their Models
AI firms buy rare books in bulk, scan them for training data, then shred the originals, with ISBNdb handling anonymous orders.
- Anthropic executed a covert 2024 operation named "Project Panama," purchasing, scanning, and destroying millions of physical books to train its Claude chatbot using hydraulic machines that sliced spines before digitizing, leaving no physical copies behind.
- AI firms target pre-2022 texts because they predate AI-generated content, offering curated knowledge that web crawls cannot replicate. A federal judge ruled this "destructive scanning" constitutes fair use, as it merely replaces print with searchable digital versions.
- ISBNdb facilitates bulk orders of up to one million books for clients, often using NDAs to maintain anonymity. Booksellers report that rare, out-of-print titles with few surviving copies are frequently processed in this pipeline for data extraction.
- Critics argue the practice destroys irreplaceable human history, noting that rare 18th-century texts are sacrificed to improve marketing emails. ISBNdb states that "AI company destroys two million books" is not a headline that generates sympathy.
- The battle over training data has shifted from courtrooms to book warehouses, affecting availability for collectors and students. As institutional buyers with immense purchasing power dominate, rare titles are permanently removed from circulation, leaving no replacement copies.
39 Articles
39 Articles
AI companies buy up antique and rare books to train their models with their content, then shred and destroy the volumes.
AI firms shred millions of print books to train models
After tainting natural resources in the name of AI, tech giants have now set their sight on books to train their new AI models. Companies like Anthropic are buying and destroying millions of second-hand antique books through mediators, slicing their spines using industry grade hydrolic press, to then feed into high-speed scanners to train AI models. To avoid training their models on AI-generated content—often referred to as “AI slop”—that has pr…
AI's hunger for data is now consuming rare books and booksellers are cashing in
Artificial intelligence firms are purchasing vast quantities of physical books. These books are scanned to train large language models and then destroyed. Rare and out-of-print titles are particularly sought after for this purpose. Booksellers report a significant surge in demand for these older publications. This practice raises concerns about the loss of scarce literary works.
Coverage Details
Bias Distribution
- 56% of the sources lean Right
Factuality
To view factuality data please Upgrade to Premium
