In this conversation, Christopher Parsons and Speaker B examine how AI‑driven search tools (e.g., UpToDate, ChatGPT, Google, YouTube) are reshaping what it means to be a medical expert. They contrast the normative view—medical education and continuing‑medical‑education (CME) still demand memorization of facts such as drug interactions and metabolic pathways—with the descriptive reality that clinicians increasingly rely on searchable databases and AI assistants for rare or “long‑tail” clinical problems. The dialogue highlights that current CME credits are earned through lectures, article reading, and quizzes, reinforcing a declarative‑knowledge paradigm, while real‑world practice shows doctors pulling up videos or AI answers to perform tasks they haven’t done since training. Parsons argues that curricula should shift from rote memorization toward systems thinking, collaboration, and teamwork skills that better align with how care is delivered in an era of perfect information.
Chapters
Listening to a podcast like 3 or 4 months ago, so I've forgotten the name of the product. I'm sure you know what the product I'm about to talk about is that makes all of the literature very, very searchable by doctors. And so one of the kind of core arguments was, look, I guess it raises questions about what it means to be an expert because in healthcare, as in architecture, engineering, you've got a certain number of continuing education hours you have to do. You want to stay up to date with the latest. But the volume of research that's getting published just accelerates every year, and no human could stay on top of just within their specialty, let alone a generalist who has to cut across specialties. And so these AI tools are able to— like, you start describing symptoms and you can kind of cut through the literature to get better answers. But what does it— what does that mean to be an expert now? Are you even trying to keep up with continuing education in the same way? Are you just relying on Searching when it's, when the thing shows up? What are you hearing from people?
Yeah, there's the normative answer and the descriptive answer, right? Which is normatively, we still seem to be expecting everybody to know this stuff in their heads. If you go to medical school classrooms right now, with very few exceptions, they're still testing students on drug interactions, right? Drug interactions is one of the clearest use cases for AI or even just computer-aided database stuff. This drug and that drug in these doses cannot go together. It's pure synthesis, right?
It's deterministic. It's like you don't even need it.
Exactly. Yeah. Right. And yet we've had tools like UpToDate, right? Which is this searchable database of all the literature, right? We've got all sorts of stuff and we haven't seen this shift towards changing the, the educational paradigm. We're still sort of treating the expertise that needs to get developed as the same expertise from 25, 30 years ago. This has been sort of a personal crusade of mine to try to get like drug interactions and the Krebs cycle, which is sort of how we metabolize something, right? I don't even know, right? It comes up very rarely and yet every medical student tells me they spend months learning this thing. All this sort of AI perfect information should be out and what should be in is how do we collaborate? How do we work in big systems? How do we do all the sort of teaming, organizational kind of stuff, you know, which admittedly I'm biased 'cause I study those things and I think they're really relevant for how we deliver care. And yet curricular-wise, we are not seeing systems stuff being taught and we still continue to see these sort of essentially rote memorization of basic facts. So that's on the normative side, right? Normatively, I, I don't know that we've changed all that much.
And you think that's also true with continuing education for a 20-year physician? It's still, read things, get it into your head.
Absolutely. When you look at the things that people get awarded continuing medical education credits for, it's attending lectures, it's reading articles in one of your field's journals and submitting a little questionnaire at the end. It's getting emailed questions and responding to them. It's just quizzes, right? It is declarative knowledge. And again, this may be back to this issue that normatively we don't know another way of thinking about what it means to be an expert, or certainly a way of assessing someone's expertise that is not in this sort of declarative knowledge paradigm. So we may just be stuck in an old way of thinking about it. Descriptively, I mean, what you see actually happening is doctors are using ChatGPT, doctors are using Google and YouTube, and they have been for quite a while, right? Because again, you see certain things rarely enough and there are good resources that you may have somebody who is covering, right? Let's say they're covering in an emergency department or something like that, or they're covering in training. It's not uncommon to cover in very broad services. And so yeah, you're a surgeon, but you haven't done hand surgery, right? And you've got somebody who comes in with a broken finger and you just need to splint it and set it and then they'll see the specialist tomorrow. You haven't splinted a finger since medical school. What are you gonna do? You go back to your office to get materials, quote unquote, and you pull up a YouTube video on how to splint a finger, right? I have, this is a 100% real world example that I have seen happen where somebody just goes and they look it up. And again, they need to have some baseline level of knowledge and expertise. When you and I watch that video, we don't pick up the same things that they pick up. Right. But the sort of Rolodex of full perfect information, I think has never really been the way that medicine gets practiced.
Especially long tail stuff that you don't apply very often.
Absolutely. And even the, the quick common stuff develops into that sort of tacit sense of, ah, this is right, that's wrong. Let me go check out the specifics later. Or this is life-threatening. This is not, this doesn't feel right. I'm gonna chase this down a little bit further. This I feel like I could send somebody home with. That it, it is built on that sort of base of experience that then you get the specific care pathway or the specific treatment that you want to take after you go back and double-check up to date to see what the current recommendations are. You know, they need to take this medication, how much of it? Nobody really walks around with that in their head anymore. They're going back and looking that up. So I think descriptively we're seeing a lot more uptake and, and use in that regard.
