AI Companies Are Reportedly Shredding Millions of Books After Training Models

Tom’s Hardware reports that some AI companies may be acquiring large quantities of books through intermediaries, using them as training material, and then shredding the physical copies. The alleged practice could reduce storage costs and make provenance or copyright investigations harder. However, the report does not provide a complete list of companies, procurement records, destruction manifests, or independently verified counts. The story should therefore be treated as an investigative lead rather than an established industry fact. It raises broader questions about dataset provenance, copyright compliance, intermediary accountability, and evidence preservation.