We trust doctors to heal us and AI to make medicine smarter. But what happens when the data fed to our most advanced algorithms is compromised by a hidden incentive? This series investigates a critical flaw at the heart of the AI revolution in healthcare: "upcoding." We uncover how the systematic inflation of medical billing codes, designed to maximize insurance payments, creates a distorted reality. This skewed data is then used to train the Large Language Models (LLMs) that doctors are increasingly relying on for clinical advice. The result is a dangerous feedback loop where AI may learn to associate higher payments with sicker patients, leading to biased recommendations, flawed diagnoses, and a new generation of medical errors. "Code Gray" explores whether we can fix the data before the algorithm makes the wrong call.