Digital Patient Podcast

TDP 176: Stanford's Dr. Mike Pfeffer: Why we Shouldn't Always Want "Human in the Loop" for AI, Why AI Prediction Models are Only as Useful as the Problems they Solve, and Where AI Scribes Must Evolve

April 24, 2025
By
seamless

Subscribe on: RSS | SPOTIFY | APPLE PODCAST | GOOGLE | BREAKER | ANCHOR

On this episode of The Digital Patient, Dr. Joshua Liu, Co-founder & CEO of SeamlessMD, and colleague, Alan Sardana, chat with Dr. Michael Pfeffer, Chief Information and Digital Officer at Stanford Health Care and Stanford University School of Medicine, and Associate Dean at Stanford University School of Medicine, about "Why we Shouldn't Always Want "Human in the Loop" for AI, Why Prediction Models are Only as Useful as the Problems they Solve, Where AI Scribes Must Evolve, and more..." Click the play button to listen or read the show notes below.

Audio:

Guest(s):

  • Dr. Michael Pfeffer, Chief Information and Digital Officer at Stanford Health Care and Stanford University School of Medicine, and Associate Dean at Stanford University School of Medicine
  • Dr. Joshua Liu, Co-founder & CEO at SeamlessMD

Episode 176 - Show Notes:

[00:00:07] Episode preview

[00:03:19] From Chemical Engineering & Music to Internal Medicine & Informatics

[00:07:06] Don’t Start with Prediction. Start with the Problem.

[00:09:48] Academic Research vs. Real-World Deployment

[00:12:01] Why “Human in the Loop” is a Broken Concept

[00:15:55] The Threshold for Rolling Out Imperfect AI

[00:20:01] Inside Stanford’s HIPAA-Safe GenAI Playground

[00:23:49] Roleplaying OSCEs with AI: The AI Clinical Coach

[00:26:35] How AI Will Separate the Best from the Rest—but Raise the Floor for Everyone

[00:29:54] Why “Small Data” is Finally Ready to Matter

[00:31:52] Balancing Data Collection with Privacy

[00:33:37] Beyond the Note: Where AI Scribes Must Evolve Next

[00:35:26] What Clinician Adoption Really Looks Like

[00:36:14] The Underrated Key to AI Scribe Adoption: Training

Fast 5 Lightning Round:

  1. What is your favorite book or book you’ve gifted the most?
    The Friction Project by Bob Sutton and Huggy Rao.
  2. If you could instantly master any skill, what would it be?
    Singing.
  3. Would you rather have Super strength, super speed, or the ability to read people’s minds?
    Super speed.
  4. What is something in healthcare you believe that others might find insane?
    The Payment & Insurance System
  5. If you could go back in time to any event or moment, when would it be?
    The origin of humans.

The Digital Patient has been recognized as Feedspot's #1 Patient Engagement Podcast of 2025. Thank you to our listeners for making this happen!

TDP 176: Stanford's Dr. Mike Pfeffer: Why we Shouldn't Always Want "Human in the Loop" for AI, Why AI Prediction Models are Only as Useful as the Problems they Solve, and Where AI Scribes Must Evolve

Posted by:
seamless
on
April 24, 2025

Subscribe on: RSS | SPOTIFY | APPLE PODCAST | GOOGLE | BREAKER | ANCHOR

On this episode of The Digital Patient, Dr. Joshua Liu, Co-founder & CEO of SeamlessMD, and colleague, Alan Sardana, chat with Dr. Michael Pfeffer, Chief Information and Digital Officer at Stanford Health Care and Stanford University School of Medicine, and Associate Dean at Stanford University School of Medicine, about "Why we Shouldn't Always Want "Human in the Loop" for AI, Why Prediction Models are Only as Useful as the Problems they Solve, Where AI Scribes Must Evolve, and more..." Click the play button to listen or read the show notes below.

Audio:

Guest(s):

  • Dr. Michael Pfeffer, Chief Information and Digital Officer at Stanford Health Care and Stanford University School of Medicine, and Associate Dean at Stanford University School of Medicine
  • Dr. Joshua Liu, Co-founder & CEO at SeamlessMD

Episode 176 - Show Notes:

[00:00:07] Episode preview

[00:03:19] From Chemical Engineering & Music to Internal Medicine & Informatics

[00:07:06] Don’t Start with Prediction. Start with the Problem.

[00:09:48] Academic Research vs. Real-World Deployment

[00:12:01] Why “Human in the Loop” is a Broken Concept

[00:15:55] The Threshold for Rolling Out Imperfect AI

[00:20:01] Inside Stanford’s HIPAA-Safe GenAI Playground

[00:23:49] Roleplaying OSCEs with AI: The AI Clinical Coach

[00:26:35] How AI Will Separate the Best from the Rest—but Raise the Floor for Everyone

[00:29:54] Why “Small Data” is Finally Ready to Matter

[00:31:52] Balancing Data Collection with Privacy

[00:33:37] Beyond the Note: Where AI Scribes Must Evolve Next

[00:35:26] What Clinician Adoption Really Looks Like

[00:36:14] The Underrated Key to AI Scribe Adoption: Training

Fast 5 Lightning Round:

  1. What is your favorite book or book you’ve gifted the most?
    The Friction Project by Bob Sutton and Huggy Rao.
  2. If you could instantly master any skill, what would it be?
    Singing.
  3. Would you rather have Super strength, super speed, or the ability to read people’s minds?
    Super speed.
  4. What is something in healthcare you believe that others might find insane?
    The Payment & Insurance System
  5. If you could go back in time to any event or moment, when would it be?
    The origin of humans.

The Digital Patient has been recognized as Feedspot's #1 Patient Engagement Podcast of 2025. Thank you to our listeners for making this happen!

Recent news from SeamlessMD

TDP 180: Endeavor Health’s Justin Brueck: The Business Case for AI Scribes, Rethinking Care Delivery from First Principles to Solve for The Silver Tsunami, and Embedding Genomics into Patient Care
June 12, 2025

TDP 180: Endeavor Health’s Justin Brueck: The Business Case for AI Scribes, Rethinking Care Delivery from First Principles to Solve for The Silver Tsunami, and Embedding Genomics into Patient Care

Learn More
TDP 179: Froedtert Health’s Dr. Brad Crotty: What Startups Get Wrong Working With Health Systems, Using Qualitative Data to Understand why Tech Isn’t Working, and Should Health Systems Still Have Venture Arms in 2025?
June 5, 2025

TDP 179: Froedtert Health’s Dr. Brad Crotty: What Startups Get Wrong Working With Health Systems, Using Qualitative Data to Understand why Tech Isn’t Working, and Should Health Systems Still Have Venture Arms in 2025?

Learn More
TDP 178: UVA Health’s CMIO Dr. Andrew Markowski: Why AI Models Succeed or Fail, the Standardization Myth, and How to Increase Your Sphere of Influence For Better Change Management
May 30, 2025

TDP 178: UVA Health’s CMIO Dr. Andrew Markowski: Why AI Models Succeed or Fail, the Standardization Myth, and How to Increase Your Sphere of Influence For Better Change Management

Learn More