AI companies buying books to scan and destroy, 404 Media investigation finds

A concerning trend has emerged where artificial intelligence companies are allegedly purchasing large quantities of books from used and independent bookstores. Investigations by 404 Media suggest these acquisitions are not for resale or archival purposes, but rather for a more destructive end.

Bookstore owners have reported unusually large orders, leading to suspicions that the books are being acquired specifically to be scanned. This process is believed to be part of training AI models, which require vast amounts of text data to learn and develop. The sheer volume of these purchases has raised alarms within the bookselling community.

“

Some used and independent bookstores have reported unexpectedly large book orders and suspect that the books are being scanned to train AI models and then destroyed.

Adding to the concern, the investigation indicates that after the books are scanned for AI training, they are subsequently destroyed. This practice raises questions about the ethics of using copyrighted material for AI development without proper compensation or consideration for the physical books themselves. A journalist from 404 Media shared insights into the findings, highlighting the potential impact on the literary ecosystem.

“

As part of a new investigative report, 404 Media put a tracking device in a book to see where it'd end up.