AI companies are secretly buying, scanning, and destroying millions of physical books to train their models, locking knowledge inside private servers. Anthropic’s Project Panama was exposed in a $1.5 billion copyright settlement. In early 2024, they launched a highly confidential project, spending tens of millions of dollars purchasing millions of paper books, scanning them, training its Claude LLM, and then destroying them all.

The motive is to keep competitors from accessing untouched pre‑2022 data, avoid legal risks, and reduce scanning costs. By monopolizing digital copies, AI firms keep knowledge on private servers forever. This paradox shows that while promising accessible AI, the companies dismantle the most solid carriers of human knowledge.

Anna’s Archive, the world’s largest shadow library, urges volunteers worldwide to scan and upload books, journals, newspapers, and rare texts. With 10 million volunteers, the archive could preserve invaluable wealth. Rewards include recognition, lifetime membership, and scanning fee assistance.

The race is urgent: since early 2025, AI content dominates new internet publications. If AI absorbs every human sentence, future generations may face a world where only AI‑written words survive. Shadow libraries offer a counter‑measure, ensuring humanity’s wealth remains free and accessible.