How Not to Get Killed Building AI in Healthcare

1–3 credits · Graduate

Clinical AI is advancing faster than the institutions required to adopt, regulate, reimburse, and govern it. A product that appears indispensable today may be rendered obsolete tomorrow by a new foundation model, reimbursement decision, regulatory interpretation, health-system procurement policy, or recognition that the customer, user, payer, and patient are often different people. This course will equip aspiring healthcare builders to navigate that uncertainty. It will be taught from three complementary perspectives: a Stanford medical student and emerging clinical AI entrepreneur actively confronting these questions; a successful health technology founder with more than a decade of company-building experience; and a health-system and policy leader whose career has included government service, health-system leadership as COO of Kaiser Permanente, and entrepreneurship as a cofounder of Qualified Health. Through candid case discussions and conversations with leaders from AI-forward health systems, successful founders and operators, investors, and representatives from organizations such as ARPA-H, HHS, OSTP, and FDA, students will examine how technological change, clinical workflows, data access, regulation, reimbursement, liability, politics, capital markets, and incumbent incentives jointly determine what can actually be built, deployed, and scaled. Students will complete a concise applied project identifying a clinical AI opportunity and stress-testing it across technical feasibility, clinical value, workflow integration, regulation, reimbursement, business model, and financing. The project will culminate in a build-or-don't-build recommendation. This course addresses a particularly urgent and currently unmet educational need. Healthcare AI is entering an unusually consequential period in which model capabilities, regulation, reimbursement, procurement expectations, and health-system deployment practices are changing simultaneously. Students interested in building or leading clinical AI organizations need more than high-level discussions of digital health, innovation, or AI's potential. They need practical preparation for the operational decisions that determine whether a clinical AI product can be integrated into care, withstand evidentiary and regulatory scrutiny, secure reimbursement, manage liability, and become a durable organization.