Freakonomics Radio Episode 661: “Can A.I. Save Your Life?”
Date: January 30, 2026
Host: Stephen J. Dubner
Guests: Dr. Robert Wachter (UCSF), Dr. Pierre Elias (Columbia University)
Episode Overview
In the final installment of the “Guide to Getting Better” series, Freakonomics Radio dives deep into the future of artificial intelligence (AI) in healthcare. Host Stephen Dubner is joined by Dr. Robert Wachter (Chair, Dept. of Medicine, UCSF and author of A Giant Leap: How AI Is Transforming Healthcare) and Dr. Pierre Elias (Cardiologist, Columbia, Medical Director for AI at New York Presbyterian Hospital). Together, they explore why healthcare delivery is lagging behind the massive technological advances in medicine, where AI is making an immediate difference, the promise and pitfalls of its integration, and how it might change the physician’s role—and patient care—for years to come.
Key Discussion Points & Insights
1. The Paradox of Modern Healthcare: Technology Inside, but Sloppy Outside (00:01–07:30)
- Wachter’s Background: As a hospitalist and department chair, Dr. Wachter created the hospital medicine field. He explains that while hospitals are technologically advanced in specialized departments (like radiology), the overall system is outdated in its workflows—still reliant on pagers and slow processes ([02:07][Dr. Elias], [03:46][Dr. Wachter]).
- Barriers to Innovation: Healthcare lags because of high fixed costs, strong incumbents, complicated economics/regulations, and a culture that is wary of disruptive changes ([06:30][Dr. Wachter]).
- Gradual, Then Sudden Change: “How does a man go bankrupt? Two ways. Gradually, then suddenly. So that’s us.” ([07:56][Dr. Wachter])
2. History Lessons: 50 Years of AI Hopes and Frustrations (09:41–17:55)
- Early AI Attempts: Initial AI tools (1970s–1980s) fell apart due to simplicity (“if/then” logic) and poor data availability (mostly on paper).
- “We’re not making the mistake this time of starting with the hardest problem” ([12:28][Dr. Wachter])
- Electronic Health Records (EHRs): Massive digitization occurred by 2016, but just shifting to digital records often made life harder for doctors (more paperwork, less face time with patients).
- “It made the lives of both patients and doctors harder. Just digitizing the record helped in certain ways … [but] wasn’t enough to transform medicine.” ([07:56][Dr. Wachter])
- Notable moment: A child’s drawing in JAMA depicted a physician with his back to the patient, typing—an unintended consequence of digital health records ([16:15][Dr. Wachter]).
3. AI’s First Major Win: Digital Scribes and Reducing “Pajama Time” (19:16–24:10)
- AI Scribes: The adoption of AI scribes allows doctors to record conversations and automatically generate patient notes, weaving relevant information together and sparing doctors from endless data entry ([19:16][Dr. Wachter]).
- “This has really been the first AI tool that took medicine by storm, and I think quite smartly … because it’s an easy win.” ([19:48][Dr. Wachter])
- Secure Integration: AI tools at institutions like UCSF are run inside secure firewalls, not via public tools like ChatGPT ([21:38][Dr. Wachter]).
- Risk of AI Errors vs. Human Errors: Human fallibility (e.g., misreading abbreviations) is not rare. AI can help avoid such mistakes.
- “Even if this chart summarization is imperfect ... don’t compare me to the Almighty, compare me to the alternative.” ([22:13][Dr. Wachter])
- Impact: Genuine reduction in after-hours work (“pajama time”) and increased openness among clinicians to further AI use ([24:08][Dr. Wachter]).
4. AI Moving into Clinical Diagnosis: From Theory to Life-Saving Practice (26:48–38:04)
- AI for Undiagnosed Disease: Dr. Pierre Elias recounts a case where late diagnosis of valvular heart disease contributed to a patient’s death and describes how AI can detect structural heart disease via common, inexpensive EKGs—something previously thought impossible ([27:32][Dr. Elias], [30:18][Dr. Elias]).
- AI Outperforms Cardiologists: In trials, AI models predicted structural heart disease from EKGs better than expert cardiologists: 78% for AI vs. 64% for doctors ([34:29][Dr. Elias]).
- “The cardiologists were as far away from random chance as they were from the AI model … in being able to predict …” ([34:28][Dr. Dubner])
- Deployment: The largest cardiovascular AI screening trial (“CACTUS”) is underway, with automatic AI screening in emergency departments ([35:10][Dr. Elias]).
- Reluctance to Adopt: Many clinicians’ resistance to AI is shaped by past “tech” promises that only made their work harder ([36:49][Dr. Elias]).
- AI as a Collaborative Tool, Not a Replacement: “1,000%. No one was talking about this … It’s been 10 years of people saying, ‘No, that’s not possible.’” ([37:32][Dr. Elias])
5. Incumbents, Monopolies, and Market Tensions (39:52–46:36)
- Epic’s Dominance: Epic, founded by Judy Faulkner, now holds records for 325 million Americans and insists on a closed, integrated system ([40:31][Dr. Wachter]).
- “Generally, monopolies are bad … I think it’s going to become more true in the coming years because of AI.” ([41:40][Dr. Wachter])
- Third-Party Innovation vs. Incumbents: Epic’s closed architecture may stifle specialized innovation; federal regulators are pushing for more openness ([43:05][Dr. Wachter]).
- Why Big Tech Fails: Tech giants like Google, Amazon, and Microsoft repeatedly falter in healthcare due to its local complexity and entrenched workflows ([43:21][Dr. Wachter]).
- Epic’s Future: Epic is committed to remaining private, as stipulated in Faulkner’s will ([45:18][Dr. Wachter]).
6. The Challenge of Regulation and the “Wicked Problem” (46:36–49:17, 58:43–61:56)
- Regulatory Gaps: Existing frameworks (FDA, etc.) aren’t designed for adaptive, context-dependent AI tools that continuously evolve ([46:42][Dr. Wachter]).
- “Our existing structures … are not fit for purpose for this tool.” ([46:42][Dr. Wachter])
- Need for Nuanced Regulation: Wachter advocates for a “light regulatory touch” to encourage innovation while being mindful of patient safety ([49:17][Dr. Wachter]).
- Human and System Factors: Technological transformation is as much about culture, economics, and politics as it is about code and chips ([58:43][Dr. Dubner]).
- “The tech can work. But then we get those layers of people ... It’s like what some people like to call a wicked problem.” ([58:43][Dr. Dubner])
- Wachter remains optimistic: “Everybody recognizes that we are in desperate need of reform in healthcare ... At least for the foreseeable future, this is gonna help me do my job.” ([59:59][Dr. Wachter])
7. AI’s Long-Term Impact: The Physician’s Role, Deskilling, and Empathy (49:36–58:23)
- AI as Co-Pilot: In the next decade, AI is expected to assist with decision support and care navigation rather than directly replacing clinicians ([55:03][Dr. Wachter]).
- “The fundamental question of AI in healthcare is not creating my note or reviewing my chart. It is computerized decision support.” ([55:03][Dr. Wachter])
- Tech companies may declare themselves “co-pilots” but “they obviously do want to replace the doctor.” ([56:59][Dr. Wachter])
- Deskilling Risks: Relying on AI may erode clinicians' skills, as seen when GI doctors performed worse after using AI tools that were later removed ([52:56][Dr. Wachter]).
- Empathy & Communication: Early studies surprisingly found patients responded to AI-generated messages as (or more) empathetic than those from human doctors ([57:05][Dr. Wachter]).
- “Of course, the AI has no empathy, but it can fake it really, really well.” ([57:05][Dr. Wachter])
- Patients Using AI: AI helps empower patients to research their conditions, which in many cases is seen as beneficial by physicians ([57:37][Dr. Dubner], [58:23][Dr. Wachter]).
8. A Vision for the Future: Will AI Save Your Life? (49:36–61:56)
- Access & Quality: AI could standardize exceptional care, close gaps, and optimize cost, but success depends on overcoming human, organizational, and cultural obstacles.
- The Reality Check: There are many constituencies with divergent incentives—“a wicked problem”—but the overwhelming dysfunction and urgent need for reform in US healthcare make AI’s promise uniquely attractive.
Notable Quotes and Memorable Moments
-
On Gradually/Suddenly Revolution:
“The quote I like is Hemingway’s... ‘How does a man go bankrupt? Two ways. Gradually, then suddenly.’ So that’s us.”
— Dr. Robert Wachter ([07:56]) -
On Digital Health Records’ Downside:
“No doctor was happy about being transformed into a data entry clerk. And patients noticed it … their doctor had the head down, typing away.”
— Dr. Robert Wachter ([16:15]) -
On Early AI and Humility:
“One of the early lessons was: Don’t start on the hardest problem with the highest stakes where if you get it wrong, you can kill somebody.”
— Dr. Robert Wachter ([12:28]) -
On AI Outperforming Doctors:
“We had 13 cardiologists look at 3,000 EKGs. … The cardiologists were at 64%. [AI] was 78%.”
— Dr. Pierre Elias ([34:29]) -
On AI’s Role in Avoiding Missed Diagnoses:
“Human intelligence is quite fallible … It’s not a matter of someone being not intelligent. There’s just no way to get the work done …”
— Dr. Robert Wachter ([23:57]) -
On the Risk of Deskilling:
“These were doctors with an average of 10 years of experience doing this procedure. So in just three months of exposure to this AI crutch, they got less good…”
— Dr. Robert Wachter ([52:56]) -
On Monopolies and Innovation:
“There’s just no way that one company sitting on farmland 10 miles from Madison can possibly do that. Their ambition is to do that. I think the world is a better place if this is more open...”
— Dr. Robert Wachter ([45:30]) -
On AI and Empathy:
“Of course, the AI has no empathy, but it can fake it really, really well.”
— Dr. Robert Wachter ([57:05]) -
On the Broader Challenge:
“The tech can work. But then we get those layers of people who may feel their realms are being infringed upon … It’s a wicked problem.”
— Stephen Dubner ([58:43]) -
On Optimism Amid Complexity:
“At least for the foreseeable future, this is gonna help me do my job. … That leaves me optimistic.”
— Dr. Robert Wachter ([59:59])
Timestamps for Major Segments
- AI vs. Workflow Dysfunction Intro: 00:01 – 07:30
- AI’s Past in Healthcare: 09:41 – 17:55
- AI Scribes & Less Physician Burnout: 19:16 – 24:10
- AI for Disease Detection and Screening (Dr. Elias): 26:48 – 38:04
- The Market: Epic, Monopolies & Innovation: 39:52 – 46:36
- Regulation and the Wicked Problem: 46:36 – 49:17, 58:43 – 61:56
- The Evolving Physician and Deskilling: 49:36 – 58:23
Overall Tone
The discussion is intelligent, candid, and occasionally wry. Wachter and Elias blend enthusiasm and wariness: they’re excited about AI’s potential (especially now that practical wins are occurring), but they’re acutely aware of medicine’s complexity, inertia, and the risks of both technological failure and dehumanized care. Dubner facilitates the exploration with characteristic curiosity and humor.
Summary in One Sentence:
Freakonomics Radio provides an in-depth, illuminating exploration of how AI could finally bridge the gap between breathtaking medical advances and the often frustratingly backward reality of healthcare delivery—while cautioning that the hardest work is still to come, and much of it will hinge not on silicon or software, but on people, institutions, and the systems that unite (or divide) them.