AI Companies Scanning and Destroying Physical Books for Training Data
fortune · reddit · 2026-08-01
According to Fortune, AI companies are increasingly purchasing large quantities of physical books for "destructive scanning" to acquire high-quality training data.
- The Incident: A Dutch bookseller received a request for 3,000 copies of a book but dismissed it as phishing because the email didn't explain the purpose.
- Context: This highlights the lengths AI companies go to for training data. Court records previously revealed that Anthropic purchased millions of physical books, removed their spines for high-speed scanning, and discarded the originals to build a searchable digital library (internally called "Project Panama").
- Legal Stance: A federal judge previously ruled that using legally purchased books to train AI models constitutes fair use, and related claims were settled.
Related event: Anthropic Exposed for Destroying Books to Train AI(4 posts)→
More from Safety
- India's CERT-In warns AI agents are becoming attack vectors — Paimaamu · 2026-10-16
- OpenAI internal model used someone else's Sepolia wallet without authorization; sleuth traces on-chain tx — banteg · 2026-10-16
- Researchers rediscover Vercel KVM escape in 2 hours using AI agent and an unredacted screenshot clue — evilsocket · 2026-10-16
- Users report Gemini's memories can't actually be deleted even after wiping chat history — ThrowRA_Purpose2025 · 2026-10-16
- You consented, but your friends didn't: Meta's Muse sparks AI privacy debate — LuizaJarovsky · 2026-10-16
- AI agents aren't 'going rogue' — frontier labs are skipping security basics — Sam_Witteveen · 2026-10-16