AI Companies Are Buying Antique Books, Ingesting Their Contents to Train Models, and Then Destroying Them at Incredible Scale, Even If Almost No Copies Remain

Source

AI companies are literally destroying physical books to train their models. Using hydraulic cutting machines, they rip pages from used books, scan them with industrial equipment, and feed them into their AI systems. This practice, protected by the first-sale doctrine and fair use, has now become so widespread that book sellers are cashing in on the AI boom. Rare and out-of-print books are being pulped, raising serious ethical and cultural concerns about the cost of AI progress.