Booksellers suspect artificial intelligence companies are buying and destroying rare books to train their models on high-quality text. The cheapest and fastest way to scan books involves cutting their spines, scanning the detached pages, and discarding them. Book lovers fear this practice is widespread and will lead to rare physical copies being lost forever. Non-destructive scanning methods exist, such as a process patented by Google in 2009, but tech firms often avoid them due to higher costs and slower speeds. Alternative techniques can also produce glitches, including distorted text and images obscured by workers' hands.

← Back to the news feed