Should AI Replace Your Doctor?

Should AI Replace Your Doctor?

“Will AI replace doctors?” makes for a great headline, a great dinner party argument, and a great way to stress out a podcast co-host. It’s also, unfortunately, the wrong question. In a follow up to our last episode, we respond to our listeners’ feedback. We put co-host and radiologist Jonathan Mates on the spot and use radiology as our test case — the specialty that’s been at the center of this conversation longer than any other, partly because the job looks simple from the outside. You look at pictures. AI looks at pictures. What’s the problem? It turns out there are a lot of problems. We get into what radiology actually involves, why AI can be impressive and still not replace humans, and some other factors we need to consider when answering this (wrong) question. Links: Geoff Hinton on radiology [https://youtu.be/2HMPRXstSvQ]

How should you and your doctor use AI?

One in four ChatGPT users — roughly 200 million people a week — use it for health questions. About a third of them end up changing their medications or requesting referrals based on what it tells them. Is that brilliant, terrifying, or both? In this episode, we dig into the question everyone is asking themselves, whether they want to or not: How should my doctor be using AI? And how should I? We bring in Rachel Lomasky [https://www.linkedin.com/in/rlomasky?utm_source=share_via&utm_content=profile&utm_medium=member_ios] — an AI scientist and expert — to help us give both the patient perspective and a realistic view of what AI is capable of. Links: Large Language Model Performance and Clinical Reasoning Tasks [https://jamanetwork.com/journals/jamanetworkopen/fullarticle/2847679?utm_source=openevidence&utm_medium=referral] Are AI Tools Ready to Answer Patients’ Questions About Their Medical Care?  [https://jamanetwork.com/journals/jama/fullarticle/2846269?utm_source=openevidence&utm_medium=referral] AMERICA’S LARGEST CITY HOSPITAL SYSTEM READY TO START REPLACING RADIOLOGISTS WITH AI, ITS CEO SAYS [https://futurism.com/artificial-intelligence/hospital-ceo-ai-radiology]

When Your Results Go Viral

People are sharing medical results from their online portals, just like they do with college acceptances and baby gender reveals. How did we get to this point? And what are the real effects on how patients understand their medical records and how doctors document them? In this episode, we explore how technology and regulatory changes have shaped our relationship to healthcare from the inside and out. Links: Why am I Watching People Get Their Medical Results? [https://www.nytimes.com/2026/04/07/magazine/medical-records-reveal-videos.html] Why Reading Your Doctor’s Notes Can Be Painful [https://nautil.us/why-reading-your-doctors-notes-can-be-painful-257796]

Why is Your Doctor's Website so Terrible?

We’ve all had this experience: you go to your doctor or hospital’s web page to do something. Maybe you want to make an appointment. Maybe you want to refill a prescription. Maybe you just want to know when the lab opens. What should be a simple task—something you can easily do on any other website—instead turns into a frustrating ordeal that leaves you yelling: “Why?!” Why is something so simple so hard to do? Why can’t they make this better? Why is this so stupid? In this episode we dig into that very question, sharing our own experiences working with these technologies from both inside and outside of healthcare, and shedding some light on how we all got here.

Should you get that test your doctor ordered?

Doctors’ ability to see inside their patients with lab and imaging tests is one of the miracles of modern medicine. Without tests, we’re just guessing!  But the unsaid truth is that doctors overtest patients, and that may be doing more harm than good to patients, doctors, and the entire health care system. In this episode, we sit down with cardiologist Dr. Jeff Sobel to understand the forces behind overtesting and how he is taking on the problem in his system. Show Notes: 1. Overuse of diagnostic testing in healthcare: a systematic review  [https://pmc.ncbi.nlm.nih.gov/articles/PMC8685650/#s5] 2. The Landscape of Inappropriate Laboratory Testing: A 15-Year Meta-Analysis [https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0078962] 3. Why Doctors Test Too Much [https://nautil.us/why-doctors-test-too-much-1188119/] 4. Projected Lifetime Cancer Risks from Current Computed Tomography Imaging [https://jamanetwork.com/journals/jamainternalmedicine/fullarticle/2832778] 5. Getting to “No”  [https://pmc.ncbi.nlm.nih.gov/articles/PMC4090224/]