The Secret Behind Mysterious Bulk Book Orders
As tech companies race to secure high-quality AI training data, secondhand bookstores across Canada and globally are facing a wave of suspicious bulk orders. Recent copyright lawsuits and industry investigations have revealed that major artificial intelligence developers are purchasing millions of physical books, slicing off the bindings for high-speed scanning, and recycling what remains to feed their language models.
Why Tech Giants Target Pre-ChatGPT Books
The sudden interest in older paperbacks comes down to the purity of the information. Books published before the rise of ChatGPT are especially valuable as AI training data because they are free from synthetic text generated by AI. Relying too heavily on AI-generated content for future model training can lead to severe performance degradation, a phenomenon known in the tech industry as ‘model collapse.’
Global Controversy and Bookstores’ Reactions
While companies like Anthropic have been confirmed to use destructive scanning methods for millions of volumes, and Amazon has tracked rare books moving into its facilities for tech development, local bookstores remain divided. Some Canadian and international sellers welcome the chance to clear out slow-moving inventory, while others refuse orders upon realizing the books will be destroyed without ever reaching readers. As investigations continue, the line between standard recycling businesses and industrial-scale AI data harvesting remains blurry.
