We Tracked a Shipment of Rare Books. It Ended at an Amazon AI Training Facility
Summary
An investigative piece by 404 Media, cited on Simon Willison's blog, tracks a large book shipment allegedly used to feed AI training data, ending at an Amazon facility. The investigation used an AirTag to trace the path, revealing a practice of data sourcing for AI models and internal discussions about book scanning at Amazon's VGT3 facility. It raises questions about data provenance, training data ethics, and transparency in the AI industry.