Artificial intelligence companies are facing growing criticism over a practice of purchasing physical books, scanning them into digital formats, and then destroying the original copies to build datasets for training large language models. Reports indicate that companies have acquired thousands of books, including rare and out-of-print editions, before removing their bindings, scanning every page with industrial equipment, and disposing of the physical volumes. The approach is mainly used for books published before 2022, as they are considered free from AI-generated content. The debate intensified after court records revealed that some firms had launched large-scale projects aimed at converting millions of printed books into digital files. A US federal court previously ruled that digitizing legally purchased books for AI training could qualify as fair use, provided the physical copies were not duplicated or redistributed. However, copyright disputes surrounding other digital materials remain under legal review. Booksellers and preservation advocates have raised concerns that uncommon and historically valuable editions could disappear from circulation if destructive scanning continues on a large scale. Critics argue that while digitization improves AI development, the permanent loss of physical books may reduce access to rare literary and academic works, sparking a broader debate over balancing technological progress with cultural preservation.
Read more here
https://www.yahoo.com/news/science/articles/vanishing-page-ai-firms-scan-160042724.html?guccounter=1
