Podcast Summary: "How is AI Changing Your Relationship With Your Doctor?"

Is Business Broken?
Host: Kurt Nickish (Questrom School of Business, BU)
Date: April 9, 2026

Episode Overview

In this episode, the panel explores how artificial intelligence (AI) is transforming healthcare, especially the doctor-patient relationship. Building on the theme “Is Business Broken?” the discussion examines AI’s impact on diagnosis, empathy, trust, and the business and ethical challenges involved in deploying AI in medicine. The conversation features expert insights from practicing physicians, AI researchers, and healthcare policy scholars.

Key Discussion Points and Insights

1. AI Surpassing Human Performance in Medical Imaging

• Eric Topol highlights that AI now outperforms humans in interpreting medical images, with randomized trials in mammography showing "remarkably increased breast cancer detection" (02:36).
• AI’s "superhuman eyes" can detect patterns doctors might miss, though these capabilities are not yet fully realized in clinical practice.

2. AI in Physicians' Daily Work

• Topol explains that speech-to-text tools now auto-generate clinical notes, reducing clerical time and improving accuracy in lab/test interpretation (03:32).
• Quote: “Conversations in the clinic [are] made into... notes and all the downstream functions to reduce the amount of data clerk time...” (03:32)

3. Unexpected AI Empathy

• A surprising theme is AI’s ability to exhibit perceived empathy. “In 13 studies... 12 of the 13 studies showed empathy was better with the AI than by the physician.” (04:06, B/Topol)
• The explanation? AI is tireless, never rushed, and can "coach" doctors to be more empathetic by pointing out missed cues in patient interactions (05:06).

4. Human-AI Collaboration and the Need for Time

• Physicians are often too rushed to express empathy; AI can flag missed opportunities and foster better communication (05:06).
• Systems combining human and artificial intelligence are expected to provide optimal care, but this relies on methodical integration and overcoming biases (06:15).

5. Data, Bias, and Reliability in AI Healthcare

• Michael Hanson (Microsoft) discusses AI’s role in not only interpreting but improving medical images, even cleaning up poor quality scans. This could benefit both developed and developing regions by making diagnostics more accessible (06:27–08:05).
• On Data Bias: Comprehensive, diverse datasets are crucial. “Unfortunately, it's not as simple as just adding more data… [datasets] are inherently biased” (09:17).
• Evaluation and continuous monitoring for bias is as important as model development (09:17–10:08).

6. Patient Trust in AI

• Patients generally trust healthcare when a doctor supervises the AI, but there's resistance when AI makes solo decisions (11:09–12:03).
• “Where we do see a reluctance... is when the AI system makes a decision without human oversight.” — Kerry Morwech (11:09)

7. The Messiness of Direct-to-Consumer AI Health Advice

• Patients increasingly use chatbots like ChatGPT for advice, sometimes receiving alarming or inaccurate information (12:37–13:36).
• Notable Moment: Kerry shares an experience where AI incorrectly alarmed her over minor lab deviations (13:04–13:36).
• Sunny Jha points out that context is critical—AI often misses nuances a physician would catch (13:36–14:26).

8. AI Hallucinations, Wrong Advice, and Deepfakes

• Stories where ChatGPT Health downplayed emergencies, advising people to stay home when they should seek urgent care (Mount Sinai study, 15:15).
• Concerns about "deepfake" digital doctor imposters eroding trust and misrepresenting real medical professionals (17:10–18:15).

9. Changing the Status and Meaning of “Doctor”

• Jha: AI complicates who can claim medical authority and increases risks around digital identity, bias, and liability (18:24–20:17).
• Doctors remain liable for their advice, unlike influencers or tech firms.

10. Systemic and Business Model Challenges

• The drive to scale and cut costs may overshadow the goal of better care (23:47).
• Administrative priorities risk eroding the patient-doctor bond unless clinicians assert the need for AI to restore—not further erode—face time (23:47–25:07).
• Topol: “It's going to take the medical professionals to stand up to these bean counters... They just understand one thing, which is revenue.” (24:22)

11. Liability and Professional Resistance

• AI introduction may shift legal liability from doctors to device manufacturers or health systems, potentially lowering malpractice costs but raising new economic and ethical issues (26:12).

Notable Quotes & Memorable Moments

• Eric Topol (On AI empathy):
  “The AI has no idea what empathy is, but it transmits it. And in 13 studies... empathy was better with the AI than by the physician.” (04:06)
• Sunny Jha (On bias and context):
  “The context of the data is important, and I think AI kind of struggles with that if it doesn't look at things in the context of a larger situation.” (14:15)
• Michael Hanson (On patient trust):
  “I don't think that patients necessarily trust a model. I think they trust the care that they receive.” (10:17)
• Kerry Morwech (On patient receptivity):
  "Patients are perfectly fine with a physician relying on AI... Where we do see a reluctance... is when the AI system makes a decision without human oversight." (11:09)
• Sunny Jha (On deepfakes):
  “…could be a deep fake, you don't know who it came from, who programmed the algorithm, who owns the algorithm...” (17:10)
• Eric Topol (On business interests):
  “They just understand one thing, which is revenue. So I’m worried about that...” (24:22)

Timestamps for Key Segments

• AI Outperforming Humans in Imaging – 02:36
• AI and Empathy Surprises – 04:06
• Data Bias and Reliability – 09:17
• Patient Trust and Human in the Loop – 11:09–12:03
• Risks of Direct-to-Consumer AI Diagnoses – 13:04–14:26
• Mount Sinai/AI Misdiagnosis Study – 15:15
• Threat of Deepfakes & Physician Identity – 17:10–18:15
• Changing Meaning of “Doctor” & Liability – 18:24–21:13
• Business Threats: Cost vs. Care – 23:31–25:07
• Shifting Legal and Economic Risks – 26:12

Conclusions & Thematic Takeaways

• Promise and Peril: AI holds immense potential for expanding access and improving care but raises new risks around bias, errors, dehumanization, and liability.
• Trust and Transparency: Trust hinges on maintaining physician oversight, transparency about AI use, and safeguarding identity and professional standards.
• Business Model Challenge: The future impact of AI in medicine will be mediated not just by technology, but by the incentives and choices of healthcare organizations.
• Human Centered AI: For innovation to strengthen—not undermine—the patient-doctor bond, the medical community and policymakers must actively shape AI’s integration.

Final Reflection

The panel’s discussion ultimately circles back to the core question: Will AI in healthcare actually fix what’s broken, or will it entrench current problems under new digital forms? The answer depends on delicate choices at all levels of business, care delivery, and policy: Will AI be used to free up doctors’ time for empathy and deeper patient engagement, or as another tool to maximize revenue and efficiency at the cost of care quality? The coming years will reveal whether medicine’s adoption of AI is truly transformative or simply more of the status quo.