Several artificial intelligence companies are reportedly purchasing large numbers of old, rare and valuable books, scanning them for AI training purposes, and then destroying the physical copies through a process known as destructive scanning.
The report says booksellers have seen an unusual surge in orders, with some reportedly selling hundreds of books a week compared with only a few dozen previously. The materials allegedly being purchased include rare historical editions, limited publications, personal diaries, journals and other records with few or no surviving alternatives.
The reported motivation is the growing demand for high-quality human-written material to train advanced AI models. After purchasing the books, the pages are separated and rapidly scanned into digital form, while the original physical copies may become unusable.
AI companies reportedly argue that digitizing such material can preserve its intellectual content and make it available for future research and AI development.
Critics, however, warn that a digital copy cannot fully preserve the historical value of an original book, including its physical characteristics, rare editions, author signatures, handwritten annotations and marginal notes.
The issue also raises broader concerns about copyright, intellectual property and authors’ rights, particularly amid continuing debates over the use of copyrighted material to train AI systems.
Historians and preservation experts are concerned that if the trend continues, valuable literary and historical heritage could become concentrated in privately controlled digital databases while the original physical artifacts disappear permanently.
The controversy ultimately raises a difficult question: Should irreplaceable physical books be sacrificed in the pursuit of better artificial intelligence, or should humanity preserve its cultural and intellectual heritage alongside technological progress?
