What happens when you train an AI model not on language, but on the structured events of a patient's medical journey? The same thing: it gets better with scale. Seth Hain breaks down the largest structured medical event model in the world with NEJM AI.
How can a health care software company develop trustworthy clinical AI tools that help patients? Seth Hain, senior vice president of research and development at Epic, describes how his company is building foundation models that respect institutional autonomy, minimize burden, and prioritize safety. In this episode of AI Grand Rounds, he discusses scaling laws in structured medical data, cautious deployment for clinical interventions, and why understanding causality — not just correlation — is essential. This conversation with hosts Arjun Manrai, PhD, and Andrew Beam, PhD, reframes AI not as disruption, but as infrastructure for safer, more reliable care. Listen to the full episode: https://nejm.ai/ep39 #ArtificialIntelligence #AIinMedicine