Ten simple rules for teaching practical AI to health data teams

Artificial intelligence (AI) has rapidly entered the healthcare discourse, often framed through the lens of frontier models, generative systems, and transformative potential. Yet for many healthcare data teams, the practical path from curiosity to capability remains unclear. Concerns about data privacy, regulatory compliance, brittle systems, and misaligned expectations frequently stall progress before meaningful experimentation can begin. This paper presents ten simple rules for teaching practical AI to healthcare data teams, grounded in real-world data, privacy-first experimentation, and system-level thinking. Drawing on hands-on training experiences and existing educational and methodological literature, we argue that AI education in healthcare must prioritize workflows over models, systems over components, and institutional capability over isolated technical success. These rules are intended to help educators and organizational leaders move teams beyond the AI hype cycle toward building trustworthy, sustainable, and context-aware AI systems that matter.