We’ve been writing about AI companies buying rare books and destroying them in order to keep feeding more information to the AI chatbots. We learned a new word for it too, destructive scanning.

At the The Guardian, last month, Kathryn James, a rare books librarian, offered a bit more background about the why part:

Destructive scanning was Anthropic’s solution to a problem: in order to “train” Claude, Anthropic had to procure a large, high-quality language dataset, preferably one created before 2022 and the corrupting influence of generative AI on contemporary text. Claude needed as many language combinations as possible, to improve its ability to predict language outcomes. Anthropic needed data, lots of it, of very high quality. Books, as it happens, remain one of the best sources for complex, high-quality, long-form text. As the court’s decision relates, Anthropic hoped that books’ “well-curated facts, well-organized analyses, and captivating fictional narratives” would help “Claude write as accurately and as compellingly as Authors”.

“Why is Anthropic destroying books?,” August 5, 2026

But why destroy the books?

Anthropic chose to use pirated sources instead of dealing with copyright issues. True, it then ended up having to fork over billions to authors in the Anthropic settlement. But the authors compensated are only the ones who are alive and signed up.

Meanwhile, the dead authors are simply forgotten and their works are destroyed. As James notes, “ There is no endangered list for printed works, and little regulation of what might constitute survival of the rare or unique.”

Many old authors will now survive only as the garble of a chatbot, as lost to history as the works of countless ancient authors. Were those ancient authors less valuable than the ones whose works did survive? How could we know? Survivals are often a matter of chance.

Andrew McDiarmid is right to call for more action to preserve these books from the bots’ maws. Even if it means that students must write their own essays (sarc).