A lesson from Yale's Introduction to Medical Software course examines how regulators worldwide are responding to artificial intelligence and machine learning in medical devices. The lecture compares regulatory guidance from the FDA, Germany, and China, covering the AI life cycle, data management practices, and validation requirements for self-updating algorithms. It draws out the distinction between interpretable and explainable AI, tying the latter to GDPR's right to an explanation, and argues that data quality, not code, has become the central concern in developing and regulating machine learning models. The lecture also covers testing strategies, weighing prospective against retrospective trials and the value of independent evaluation before a model reaches clinical use. Aimed at students learning how medical software gets approved, it treats regulation as a technical problem as much as a legal one.