Home / Technology / AI's Secret Appetite: Books Pulped for Data?

AI's Secret Appetite: Books Pulped for Data?

Summary

  • Unusual bulk orders of secondhand books are being sent to distant warehouses.
  • A US court ruling suggests using books to train AI is not copyright violation.
  • AI companies may 'destructively scan' books, potentially pulping them for data.
AI's Secret Appetite: Books Pulped for Data?

Independent booksellers are reporting a surge in unusual, large-volume book orders, a phenomenon not seen in decades. Booksellers like Stuart Manley of Barter Books have noted that single orders can equal a week's typical sales, with shipments heading to unknown overseas warehouses.

The current suspicion is that these books are being acquired for artificial intelligence (AI) training. This theory gained traction following a 2025 US court ruling where a judge found that using purchased books to train AI was a 'transformative' use, not violating copyright law.

Further revelations emerged from unsealed court documents concerning AI firm Anthropic and its 'Project Panama.' Internal communications indicated an aim to 'destructively scan all the books in the world,' a process involving digitizing books at an industrial scale, potentially by removing their spines and recycling the pages.

While AI companies maintain they do not buy or destroy rare books, the prospect of destruction for common titles causes unease among booksellers. Experts note that the diverse subject matter of these bulk orders also suggests AI, as varied data helps improve AI models.

Copyright laws differ between the US and the UK, where explicit permission from copyright holders is generally required for such copying. This situation presents a dilemma for booksellers, balancing trade benefits with ethical concerns over potential book destruction.

Disclaimer: This story has been auto-aggregated and auto-summarised by a computer program. This story has not been edited or created by the Feedzop team.

Read more news on

Property Code: 5571