Inside an AI Start-up's Plan to Scan and Dispose of Millions of Books

In early 2024, executives at the artificial intelligence start-up Anthropic quietly accelerated an ambitious and secretive initiative. According to an internal planning document that was recently unsealed through legal filings, the company launched what they called "Project Panama." The document's description was strikingly blunt:

"Project Panama is our effort to destructively scan all the books in the world. We don't want it to be known that we are working on this."

This revelation comes from court filings that have shed light on how AI companies have been racing to obtain massive quantities of books to feed their chatbot systems. The approach involved buying, scanning, and then disposing of millions of titles — a process described as "destructive scanning." 📚

The internal documentation makes clear that secrecy was a priority for the project. Anthropic executives were aware of the sensitive nature of their efforts and deliberately sought to keep the initiative under wraps, understanding that public knowledge of the project could create complications.

Wrapping Up

The unsealed court documents reveal the extraordinary lengths AI companies have gone to in their quest for training data. Anthropic's "Project Panama" represents a striking example of how the AI industry's hunger for content has led to large-scale, covert operations involving the physical destruction of books after scanning them for data. As legal proceedings continue, these revelations raise important questions about transparency, copyright, and the ethics of how AI systems are being trained. 🤔