AI companies are reportedly shredding millions of books after using them to train AI models — tech giants outsource to middlemen to secretly buy up books for training material

An investigative report says AI companies and intermediaries have been bulk-buying secondhand books to extract text for model training.

The report links the practice to sellers such as ISBNdb, Better World Books, Alibris, Biblio, and Datamation Information Services, and to earlier Anthropic-related cases.

It alleges that destructive scanning methods can leave books shredded after digitization, raising copyright and transparency concerns.

The episode highlights the tension between AI data sourcing and the loss of physical copies of human-authored works.