AI, Prior Auth, and the Cost of Getting It Wrong at Scale

Описание: Artificial intelligence is moving fast into prior authorization and claims processing. For payers and providers, the promise is speed, efficiency, and lower administrative cost. But when these systems scale faster than governance and oversight, the consequences can be serious.

In this episode of Health/Tech Edge, Mike Moore is joined by Michelle Mello, Professor of Law at Stanford Law School and Professor of Health Policy at Stanford Medicine. Michelle is the lead author of the Health Affairs paper The AI Arms Race in Health Insurance Utilization Review: Promises of Efficiency and Risks of Supercharged Flaws.

They discuss why prior authorization has become a prime target for AI, which parts of the process are genuinely suited for automation, and where AI can quietly reinforce flawed incentives and bad decisions. The conversation also covers the reality behind “no denial without human review,” how AI-curated summaries influence human judgment, and why making prior auth cheaper to run can sometimes make the system worse rather than better.

Michelle also shares lessons from evaluating real AI tools inside a large health system, including how workflow pressure, limited transparency, and overreliance on automation create risks that leaders often underestimate.

This episode is a practical conversation for healthcare executives, operators, and technologists who are making AI decisions today.

Links
• The AI Arms Race in Health Insurance Utilization Review (Health Affairs) https://www.healthaffairs.org/doi/10....
• Stanford Healthcare Ethical Assessment Lab for AI (HEAL-AI)
https://heal-ai.stanford.edu/