EP. 167: AI AND THE BIGGEST EXPERIMENT IN MEDICINE
WITH ROBERT WACHTER, MD
A renowned leader in hospital medicine explains the difficulty of predicting the consequences of new technologies — and makes the case for AI-optimism in healthcare.
Listen Now
The electronic medical record (EMR) has become an unwelcome interloper in the exam room. Too often, patients find themselves answering questions delivered from behind a monitor by physicians hurriedly typing away. This isn’t the kind of care anyone wants — but it’s what the system demands. Thankfully, change may be on the horizon. AI scribes are now being rolled out in EMRs across the country, capable of listening to a visit, generating a clinic note, and freeing the physician to be present with their patient. But these scribes are only an opening act of a much larger experiment — one that asks not merely whether AI can redeem the medical record, but whether it can usher in something closer to a Golden Age of medicine.
Our guest on this episode is Bob Wachter, MD, professor and chair of medicine at UCSF and a leading voice in hospital medicine and administration. In 1996, he and his colleague Lee Goldman coined the term “hospitalist,” giving rise to what has become the fastest growing specialty in the history of modern medicine. He has authored over 300 articles and 6 books, including the New York Times bestseller The Digital Doctor (2015) and A Giant Leap: How AI is Transforming Healthcare and What That Means for Our Future (2026).
Over the course of our conversation, Dr. Wachter traces how a long-held fascination with systems drew him into studying medicine's digital transformation over the past 15 years — a period spanning the turbulent rollout of electronic medical records and now the arrival of AI. We explore the bumpy history of adopting powerful but general-purpose technologies, how such technologies force complex industries to reshape themselves, and humanity's humbling track record of predicting what comes next. Dr. Wachter makes the case that AI's integration into medicine constitutes the biggest experiment the field has ever undertaken, and explains why he believes it will ultimately resolve many of the health care system's deepest problems and elevate the practice of medicine itself.
-
Bob Wachter is Professor and Chair of the Department of Medicine at the University of California, San Francisco (UCSF). He has appeared on Modern Healthcare magazine’s list of the 50 most influential physician-executives in the US more than a dozen times; he was #1 on the list in 2015.
An elected member of the National Academy of Medicine, he coined the term “hospitalist” and is considered the father of the hospitalist field, the fastest-growing specialty in the history of medicine.
He is the author of five previous books, including the New York Times bestseller The Digital Doctor. His new book is A Giant Leap: How AI Is Transforming Healthcare and What That Means for Our Future.
Dr. Wachter is the past-president of the Society of Hospital Medicine, past-chair of the American Board of Internal Medicine, and an elected member of the National Academy of Medicine.
-
In this episode, you’ll hear about:
3:46 - Dr. Watcher’s path to medicine and to his study of digital transformation
8:06 - The wins and losses of the transition to electronic medical records
26:45 - Why Dr. Watcher is optimistic that AI will deliver a “golden age” for medicine
37:30 - Contending with the potential dangers of AI in the revenue-focused medical industry
50:30 - Dr. Watcher’s view on if there will always be an important place for doctors in the future
-
TDA 167 Bob Wachter final1.mp3
Henry Bair: [00:00:01] Hi, I'm Henry Bair.
Tyler Johnson: [00:00:03] And I'm Tyler Johnson.
Henry Bair: [00:00:04] And you're listening to The Doctor's Art, a podcast that explores meaning in medicine. Throughout our medical training and career, we have pondered what makes medicine meaningful. Can a stronger understanding of this meaning create better doctors? How can we build healthcare institutions that nurture the doctor patient connection? What can we learn about the human condition from accompanying our patients in times of suffering?
Tyler Johnson: [00:00:27] In seeking answers to these questions, we meet with deep thinkers working across healthcare, from doctors and nurses to patients and healthcare executives, those who have collected a career's worth of hard earned wisdom, probing the moral heart that beats at the core of medicine. We will hear stories that are by turns heartbreaking, amusing, inspiring, challenging, and enlightening. We welcome anyone curious about why doctors do what they do. Join us as we think out loud about what illness and healing can teach us about some of life's biggest questions.
Tyler Johnson: [00:01:02] The electronic medical record has become an unwelcome interloper in many physician patient interactions. Too often, patients are answering questions lobbed from behind a monitor by physicians hurriedly typing in the Amar. This kind of impersonal care isn't what patients or physicians want, but it's what the system often demands. There is hope for improvement across the country. Ai scribes are being rolled out in EMRs now. Ai can listen to the visit, generate a clinical note, and leave the physician to review and edit the note later, freeing the physician to interact with their patient face to face. These AI scribes are in early stage of a much larger experiment, an experiment that is not simply asking whether AI can redeem the EMR, but whether AI can usher in something closer to a golden age of medicine. Our guest on this episode is Dr. Bob Wachter, professor and chair of medicine at UCSF and a leading voice in hospital medicine and administration. In 1996, he and his colleague Lee Goldman coined the term hospitalist, which is now the fastest growing specialty in the history of modern medicine. He has authored over 300 articles and six books, including the New York Times best seller The Digital Doctor, and his 2026 book A Giant Leap How AI is Transforming Health Care and What That Means for Our Future. Over the course of our conversation, Doctor Wachter shares how his long held passion for studying systems led him to studying the digital transformation of medicine for the past 15 years, a time period that has included the widespread rollout of electronic medical records and now the advent of AI in medicine. We discuss the bumpy process of adopting generalized technologies such as EMRs and AI, how generalized technologies require complex industries to reform themselves around the technology, and humanity's dubious track record of predicting the consequences of new technology. Dr. Wachter describes the incorporation of AI into medicine as the biggest experiment in the history of medicine, and explains why he is optimistic that AI will resolve many of the health care systems problems and elevate the practice of medicine. Dr. Wachter, it's great to have you on. Welcome to the show.
Dr. Robert Wachter: [00:03:24] My pleasure. It's great to be here.
Tyler Johnson: [00:03:26] You know, you are one of the best known physicians. I think it's fair to say in the country, maybe in the world. Obviously, the chair of the Department of Medicine and a renowned medical program and all the rest. But before all of those things happened, how did you end up in medicine in the first place?
Dr. Robert Wachter: [00:03:42] Well, Tyler, it's an interesting story. Neither my parents went to college. I grew up on Long Island, New York. I think my folks kind of were excited about being a physician. There were a fair number of physicians in the neighborhood. I knew this because my dad, when he would go to a party, there would be a lot of doctors sometimes at the parties, and he would wear his garage door opener on his belt. I'd say, dad, what the hell are you doing? And he'd said, well, there are a lot of doctors there. They have beepers, I have nothing. And sometimes my sisters and I would be home and the garage door would be going up and down as dad was showing off his beeper. I had an affection for medicine. I met a lot of doctors growing up. I said, this seems pretty cool. What was weird for me was I was a political science major in college. I grew up in the Watergate era. I found politics and history fascinating. So the disconnect for me was I thought I wanted to be a doctor, but was really interested in systems and people and politics and money. And I wondered, is there any way to reconcile that? And then I went to college at Penn and went to med school at Penn. And I was very lucky at Penn and then at UCSF when I came for my residency. There were some early faculty who were doing research in the system. I didn't know that was possible. When I went to school, I thought if he did research in medicine, it was in gels and jeans and things that I didn't know anything about. And I said, wow, there's a way of combining my interest in how how systems work and being a doctor and, uh, been shocked that that's worked out pretty well.
Tyler Johnson: [00:05:07] How did that early interest in systems, health care systems in particular, take you from an aspiring undergraduate medical student all the way to where you are?
Dr. Robert Wachter: [00:05:19] Well, I came out to UCSF for my residency, met some mentors who also were very interested in the way the system worked. Also, when I came out to UCSF, Aids was just beginning and it exploded during my residency, and I wrote an article in the New England Journal when I was a third year resident called The Impact of Aids on Residency Training. And basically it starts with, I've seen more patients with pneumocystis than pneumococcal pneumonia and more patients with Kaposi's than breast cancer, and became very interested in how Aids patients did when they went to the ICU. I decided I wanted to be a general internist. Decided I needed additional training if I wanted to be an academic in epidemiology policy ethics. I came here to Stanford, where I was a fellow, something called a Robert Wood Johnson Clinical Scholars fellow, then came back to UCSF, joined the faculty, started as a researcher with kind of the job. You're supposed to want to have 80% protected time. I wrote a few big grants. They imploded, and I said, I'm actually not going to be very good at this if I want to stay in academia. I like writing, I like thinking and analyzing. But to be the kind of person who writes a grant after another in a paper, after another, something I admire a lot, but I think is really a hard way to make a living. I kind of decided that I thought there might be another path because I thought I'd like leadership roles. It's shockingly and sort of an act of administrative malpractice. I was made residency director at UCSF in my second year on the faculty, did that for a few years, then was asked by a new chair to be his kind of right hand clinical lead for for UCSF. And that was a decisive moment for me because he was a burly guy, a guy named Lee Goldman, who later became dean of Columbia.
Dr. Robert Wachter: [00:06:51] And Lee said to me, you're in charge of the inpatient service at UCSF. It seems like it's organized the same way it was when I was a resident. When meaning Lee was a resident 20 years ago. It's got to be some new way of thinking about how we organize hospital care. And together, we kind of cooked up the idea of this new specialty of general internists who focused on hospital care. And I wrote a New England Journal article called the The Emergence of Hospitalists in the American Health Care System, which was the coining of the term that became the fastest growing specialty in history. And I think that has become sort of my career, which is I've had been lucky enough to have a series of leadership roles that have been gratifying. Interesting. Love the people I work with, take up a decent amount of my time, but have also given me an altitude where I get to see how systems work. And then I like thinking about and writing about what to me is the most interesting issue in the system. That led me to spend years focusing on how to organize hospital care. A lot of years of focusing on patient safety. And then over the last 15 years, just obsessed with the digital transformation of medicine. Partly why it's gone so badly. And then over the last few years, sort of what happens as AI enters our world, a little bit of a Covid interlude over the last 3 or 4 years where I kind of segue to spending a lot of my energy thinking about Covid and how to explain it to professionals and to regular people.
Tyler Johnson: [00:08:06] I think that's some great framing for our discussion. That's a little bit of your personal background, but also gets us to really the heart of what I want to talk about today, which is that your new book, which follows on the heels of the book you've had written previously that was already dealing with what was then thought to be the digital transformation of health care. But now you're talking specifically about the nascent dawn of the AI era in healthcare. And the book is called A Giant Leap. It's sort of an outline of the promises and perils of AI in healthcare. Let's start in a little bit of a funny place. So you do this in your book, you go back to about the year 2008. This is something I remember well because I graduated from medical school in 2009. So as I arrived at Stanford, where I came for my residency from, I also did my medical school at Penn. When I arrived here as a resident, epic had just been fully integrated into the inpatient and outpatient systems here. And so I remember very much from my time in medical school, especially depending on the hospital paper records. And I remember that it was often a mess, right? Because it's not just that they were paper records, it's that sometimes you had places where you had your scan results in one binder and your vital signs. Different binder and your consult notes in a separate binder and your, you know, whatever your.
Dr. Robert Wachter: [00:09:23] And your O2 sat on a post-it note.
Tyler Johnson: [00:09:25] Pre rounding, which is the process of getting all the information ready and together for your own analysis, but also very importantly to present to the attending. Correct was just maddening because you couldn't find the binder and you didn't know where.
Dr. Robert Wachter: [00:09:38] I'll go back once. So you're so you're very young. I'll go back one step further. When I was a, when I was a resident at UCSF, at the VA. What you did, you went you went to the lab and in the lab was a shoebox. And in the shoebox were carbon copies of all the lab slips. And so this was the Easter egg hunt. You went on to try to find your patient's creatinine results.
Tyler Johnson: [00:09:58] And so I remember during the Obama presidency, this was like the big national push, right, was to get everybody on to electronic medical records. And intuitively, this seems like the biggest no brainer on earth, right? How could it possibly not be leaps and bounds? Better to have all of the medical information, whether in an inpatient or outpatient setting in what we use here at Stanford is epic, which is the gorilla in the room for EMR, right? And certainly in some ways that is true, right? When I see a patient in my cancer clinic, I can just click through and I can see when was their as long as it was done here, which is a new address. But still I can see their latest echo results and I can look at all their lab trends and I can see all the consultant notes and. In et cetera. Et cetera. As far as that goes, it would seem like that maybe this should be the panacea that people promised that it would be back when we were getting ready to implement them. And yet you also outlined that, in fact, it is very much not a panacea. Talk a little bit about what has been delivered by the EMR has been a lot more complicated than what was hoped for.
Dr. Robert Wachter: [00:10:59] Yeah, I you know, when I'd been studying and working a lot on patient safety for the decade before the EMR became ubiquitous, and they really did sort of flip a switch in 2008, when you were finishing training, fewer than 1 in 10 American hospitals had an electronic health record by about 7 or 8 years later, fewer than 1 in 10 did not. So in a very short period of time, we went from that weird, unbelievably clunky paper world to basically a digital industry. And I think it was logical to think and hope that this would be the transformative moment in healthcare, that it would. Finally, we went digital and it would therefore make health care better and safer and cheaper and more convenient and accessible. And as you say, it solved a whole bunch of problems. It got rid of doctor's handwriting as the perennial joke, you know? And you could get your wallet, your prescription electronically to Walgreens and CVS. Cvs. That was great. Two people could see the record at the same time. The patient could have a patient portal and see their results. All good things. But what I didn't recognize, and I kind of came to understand in writing my book ten years ago, the digital doctor was just taking the data and sticking it in digital form in what essentially was a big filing cabinet, solves a few problems, doesn't solve the big problem. And the big problem is how do you take this data, make sense of it, and create and put it in useful form.
Dr. Robert Wachter: [00:12:21] That then leads to actionable steps that make care better, safer, and less expensive that it did not solve that. And what I guess surprised me, and part of the reason I wrote that first book was not only did it not solve all the problems, but it created a host of new ones. The spine of that book is a kid at UCSF who gave a 38 fold overdose of septra to an error that could not have happened if we were on paper. It was sort of. There was a digital reason why the error occurred. And then the slow motion train wreck of it was obvious. This kid. What do you mean? We're giving this kid 39 septra pills. Are you kidding me? And yet everybody said, well, the electronic health record says it is, so it must be so. The thing is smarter and more expensive than I am. So I believe. And just by dumb luck, he didn't die. You know, other things I pointed out in the book in the old, old days. I don't know if you'll remember this, but in my days at Penn, when you went on rounds, you always ended rounds by going down to the radiology department. And in the radiology department, the the guru, there was a guy named Wally Miller, and you ran the case by Wally. And Wally put up a chest film and said, what do you think's going on? It was magical.
Dr. Robert Wachter: [00:13:27] And the minute you didn't have to go to the radiology department to see your films. You know, trainees don't even know what I mean when I say film or a lightbox, the mean you didn't have to go down. People stopped going down and nobody predicted that would be so. And yet. So that book was really about all the both unanticipated consequences of digitizing the record and the fact that it did not solve the problem. But it was a really very important move for us to make because it was foundational. And so that book was a pretty grumpy book about like, why does everybody hate their EHR? Why has it not delivered on the promise? And there really were lots of promises. And it looked great on the PowerPoint slides from the consultants. And I ended that book on a very optimistic note after a pretty grumpy book. The last chapter is, I think this is going to work out, but I think we did not understand a how complex this is. B that in the history of what are called general purpose technologies, they never work out in the beginning, in part because the version 1.0 is not very good. It has to get better. But much more importantly, complex industries need to kind of remake themselves, and they never do in version 1.0. They just take this technology, they plunk it in, they turn it on, and they think we're done. And so it's going to take some time.
Dr. Robert Wachter: [00:14:41] And I think I had a premonition. It's not only going to take some time, it's going to take a technology that can do some things that the technology's a decade ago couldn't do. And the main thing was to basically be able to understand human language, because if you're trying to analyze your record, the old AI up until November 30th, 2022 could analyze your creatinine results and the EKG results and the the hemoglobin, but couldn't make begin to make sense of your note. And much of the data and insight in a medical record actually is in the form of narrative. And so we were kind of waiting for a technology that could ingest a whole lot of diverse kinds of material, not only structured data, but unstructured data, notes, images, videos, all that sort of stuff. And then had a level of intelligence beyond anything we had had in AI before in terms of its ability to not only understand that, but piece it together and come out with suggestions, predictions, recommendations. And so the first time I used ChatGPT on November 30th, 2022, a little light bulb went on and said, this is what we've been waiting for. This is the moment where this healthcare system that has truly not been transformed and markedly improved. With the first version of the technology, the HR now has at its disposal a technology that really is exactly what we need to get us to the right place.
Tyler Johnson: [00:16:05] Okay, so I want to get to those thoughts, but first, I want to talk a little bit more about the bridge that we have come across, because one of the things that you alluded to early on, it would be one thing if what had happened was that a bunch of consultants in 2005 or whatever had put together slide decks and said, hey, we've got the computing bandwidth now to have an EHR. Here's how we think it'll be designed. This is what we think it'll look like. It would be one thing if 15 years thereafter or whatever, we looked back and we said, well, look, as you put it, the EHR is a really, really fancy filing cabinet and it's an incredible filing cabinet. And it's remarkable in terms of its ability to collate information and give us access to things that were done a long time ago and make sure that everything is legible and findable. And et cetera, et cetera. But that just hasn't solved some of the bigger problems that we see on the farther end. Right? That would be one thing. But actually, what has happened, and you alluded to this earlier, is that in addition to not living up to some of the grander promises potentially made back 15 or 20 years ago, it actually has created a whole slew of other problems that most doctors, as you and I know, would say, have actually, not only have they not solved problems, but they have actually made their lives actively worse, right? And of course, the one thing that is that is on virtually every physician's mind, and that I'm sure that you deal with a whole lot in your sort of roles.
Tyler Johnson: [00:17:30] Helping to manage the system at UCSF is that everybody, every doctor knows now that you spend X number of hours actually caring for patients in the day, and then you spend some portion, whether it's 0.2, 0.4, 0.6, or sometimes it seems like more than one X carrying an effect for the medical record, right? Including doing a lot of stuff that is not clear, does anything to make patients lives better, but instead makes billing more efficient, or makes insurance companies or health care corporations or whatever, more money. So I guess the question that I'm wondering about is this what factors do you think are responsible for making it so that not only did the EHR not solve all the problems that it hoped to solve, but that actually led it to creating more problems that people didn't anticipate.
Dr. Robert Wachter: [00:18:17] It's interesting because in some ways all of the problems were anticipated, but we anticipated none of them. And, and I think that just shows that humans are really quite awful at projecting forward into a world of a new technology and what life is going to be like. Henry Ford was reputed, probably never did say, but was reputed to have said, if I asked people what they wanted, they would have said faster horses. They had absolutely no ability to see what their life might be like with a car until there was this thing called a car. And so the point is really very, very important and important for us to recognize as we embark in this new era of, of a new technology. But here's the things that were anticipated, but I did not anticipate. One is that we would stop going to radiology. We would stop going to radiology rounds. Totally predictable. If you don't need to go down, you won't go down. And yet it was like the old folks like me sort of lamented. Like those were really great. You know, the interactions we had with the radiologists were fun, were interesting. They learned, we learned. Sad that it went away. And yet the minute that you didn't have to go down those, those rounds ended. What the doctor does is really important for the system, both in terms of the quality but also the cost of care, billing, etc..
Dr. Robert Wachter: [00:19:28] In the world of paper records, nobody could make me do anything. I was scribbling on a piece of paper, and if somebody wanted to see whether I was practicing good medicine or my documentation was the right documentation to create the best bill, they had to sort of decipher my chicken scratch when the record made it back to the medical records department. You now had the capacity to look over the doctor's shoulder in real time once it was electronic and also create forcing functions to say, doctor, when you call this thing. You know, weakness. We only got paid, you know, 60 bucks. But if you call it, you know, functional hemiplegia or some unbelievably, you know, catch 22 made up term, we get paid 110 bucks. So you should do that. So and it turns out that we're being measured on quality. So if you can document that you counsel the patient about seat belts and ask them whether they felt safe at home, etc., all things that sounded good when you were sitting in a conference room, but drove the doctors crazy. And lo and behold, again, totally predictable. But none of us predicted it. The life of a doctor in the clinic was. The patient is looking at the doctor, and the doctor is looking down at the keyboard because the doctor is so busy clicking all these boxes.
Dr. Robert Wachter: [00:20:38] So the fact that the electronic health record gave the power to outside entities, whether it was your your system or the insurance company or the malpractice attorney, the ability to make you do stuff and look at what you're doing in real time and force you to do stuff, created a world where the documentation requirements became massive and physicians noticed patients noticed. The final straw really, and again, talk about unanticipated consequences that are obvious in retrospect. When epic rolled out, we didn't have. It was just for the docs and the nurses. And then a few years later, they turned on a patient portal. Great. All right. Wonderful. Democratized medicine. Patients have access. And then a few years later, the federal government passed a law that said patients can see all their lab results, their X-ray results, and in fact see their doctor's note again. Great. The problem is, there was absolutely no useful intelligence given to patients to interpret any of that information. If you think about when you go on to United Airlines website or the Bank of America website, how often do you say, I need to call somebody? Almost never. They. It solves all your problems here. They gave patients a massive amount of information, but basically no tools to do anything with that information except for one.
Dr. Robert Wachter: [00:21:47] And the tool was epic. And the other EHR vendors put a little button at the bottom of the of the screen and says, click here to send a message to your doctor. And shockingly, they being human beings with normal wants and needs said, you know, I'd love to know, like I'm having a headache, what's going on? Or I see that my magnesium is low and my EKG is abnormal. That's scary. What does that mean? Oh, let me click that little button. And so doctors now had this phenomenon we call pajama time, where now you put the kids to bed and you have two hours not only of documentation, but answering emails for which you're not being compensated. And so again, all of these things seem as I look back at them, I'm shocked at myself. It's like, of course that's going to happen. And yet I don't know a single person who anticipated the end of radiology rounds, the fact that nobody's going to be looking at a patient in the eye because they're so busy pecking away, or the fact that they're going to be spending 2 or 3 hours at night answering inbox messages from patients. I think we have to get better at that. But that I think those were some of the consequences of the electronic health record that, as you say, not only didn't deliver on what we hoped it would deliver, but actually worsened the experience of clinicians and maybe even patients.
Dr. Robert Wachter: [00:22:56] So that's kind of where we were on November 29th, 2022, the day before GPT came out. And that's partly why this book is not a grumpy book. My new book is actually quite optimistic. It's partly because the technology really does solve a lot of the problems that were created by the electronic health record and really allows us to do things that are magical. But almost more importantly, on November 29th, 2022, I wasn't sitting there on Google saying, God, Google stinks. I need a better search engine. In fact, I couldn't even imagine what a better search engine would look like until GPT came out and said, wow, this is fantastic. It's better. But do you know anybody who on November 29th, 2022, said the health care system is just perfect? I don't know a doctor, I don't know a nurse, I don't know a health care leader, and I certainly don't know a patient who said the system is delivering on all cylinders. Everybody saw the friction was frustrated. Costs are impossible. Quality access, convenience, all that sort of stuff. So part of my optimism is I think we have so much room for improvement. And these tools, I think, deliver something that will allow us to do that.
Tyler Johnson: [00:24:00] So I do a lot of teaching of medical students and residents and fellows as well. One of the things that I talked with my medical students about is that when you're a medical student, getting ready to go on to the wards, Arguably, the single most closely observed thing that you're going to be doing, unless if you're a surgeon, then I suppose maybe surgery. But if you're going to be an internist, the place where you have a moment to really shine is on rounds. You present what's called an HTTP. An HTTP is where you have a new patient who's come into the hospital you're attending, at least nominally, doesn't know about the patient yet, and you are going to present the patient's history. And then you're also going to say what you think, therefore, what you make of it and what you want to do about it, right? And when I'm teaching my students to do this, I tell them to imagine that there's a bright red line right before what we call the assessment and plan, right? So everything above that, in effect, what the medical student is doing is they are gathering information and then they're presenting the information back. It's an investigative and a reporting task. And as long as they have learned, you know, how to gather which information to gather and how to cull signal from noise and all the rest of it, most of them can do that pretty well. Then after you've presented all the information, then you give a one line summary, which is kind of your way of distilling down everything you've said and saying, here's the pertinent information.
Tyler Johnson: [00:25:22] And then you launch into what's called the assessment plan, which internists know that that's really where the substance is, right? That's where a medical student and later a resident is able to take everything, all the information that they've just gone over and say, here's what I make of it and here's what I want to do with it in terms of the patient's treatment plan. So in some ways, when I think about the difference between EHR and AI, largely what the EHR was doing or was meant to do is the stuff above the red line, right? It's largely gathering information and then presenting it in a digestible, intelligible way. But the difference with AI is that now AI, arguably right now and certainly in the future, has the potential to be inserting itself into the substance of the assessment and plan. That is to say that it's actually taking a lot of this information that the EHR provides from patients and labs and nowhere else. And then trying to suggest, what do we make of this information and what would be the best thing to do with it? Now, as you point out over and over again in your book, this is a much higher stakes thing, right? And it also feels a little bit as a doctor, I think most doctors have a little bit of a there's a little bit of sort of uneasiness, like, hey.
Dr. Robert Wachter: [00:26:40] That's my thing. I went to ten years of school to learn to do that, of course.
Tyler Johnson: [00:26:44] And so let me actually just read. So you also sort of in conjunction with the publication of your book, you published an op ed in the New York Times, which had a I'm going to review a funny name, stop worrying and let AI help save your life.
Dr. Robert Wachter: [00:26:58] Yeah. You should know that the that the paper that writes the headlines and I, I hated it, but you know, the New York Times and New York Times.
Tyler Johnson: [00:27:07] So you've gone through, for example, comparing robo taxis to healthcare implementation. And in effect, you've been making the argument that, well, of course, we should pay close attention to the problems when or the times when AI gets things wrong. We shouldn't focus so much on those that we become myopic and lose sight of the benefits. Potential benefits of AI. So then summing up, you say, I'm not arguing that we shouldn't aspire to perfection, or that AI and healthcare should receive a free pass from regulators. Ai designed to act autonomously without clinician supervision, should be closely vetted for accuracy. The same goes for AI that may be integrated into machines like CT scanners and sometimes surgical robots. Areas in which a mistake can be catastrophic, and a physician's ability to validate the results is limited. We need to ensure patients are fully informed and can consent to AI developers intended use of their personal information for patient facing AI tools in high stakes settings, such as diagnosis and psychotherapy. We also need sensible regulations to ensure accuracy and effectiveness. But then you come to sort of your take home point. But as the saying goes, don't compare me to the Almighty. Compare me to the alternative. In healthcare, the alternative is a system that fails too many patients, costs too much, and frustrates everyone it touches. Ai won't fix all of that, but it's already fixing some of it. And that's worth celebrating. Let me ask you what seems like a fair question. You said a moment ago that if I had sat you down for an interview in the first couple of years, let's say, of the EHR revolution, and had asked you to predict what's this going to be like, that you probably, as many people did, would have predicted a lot of great ways that the EHR would transform medicine without predicting these things, which you now say were completely foreseeable and yet almost entirely unforeseen.
Tyler Johnson: [00:29:03] So I'm curious, why do you now think? Because one of the last chapters of the book. So you go through sort of all of these different use cases and whatever both ways that things are already being used and ways things that might be used in the future. And then at the very end, you have a chapter that is on your predictions for how AI will be integrated into healthcare over the next decades. And in that chapter, multiple times, you use some version of the phrase Golden age. Do you think AI is going to help usher in a golden age of practicing medicine? And in effect, I'm overly summarizing your argument, but very shorthand version of your argument is that you think AI will relieve physicians of many of the things that they don't like doing anyway, including, ironically relieving them of many of the functions that were added to their jobs because of a result of the way that EHR didn't go the way that we wanted to. It will relieve them a lot of a lot of the kind of administrative fluff and allow them to focus on sort of, for lack of better words, kind of the heart of medicine, But I guess the question that I have is, what is it precisely that gives you the confidence that this is going to usher in a golden age, when the last thing that we thought was going to usher in a golden age has had all of these foreseeable and yet unforeseen consequences.
Dr. Robert Wachter: [00:30:17] That's an absolutely fair question. I think the most honest answer is I'm not sure, because I think we are now in the middle of the greatest experiment in the history of medicine. We are basically transforming the nature of what we do, how we interact with patients, how patients interact with their own medical care, some of which will allow them and permit them to bypass the medical system as we currently know it. And it could go off the rails in all sorts of ways. I think I, and we collectively are a little smarter and less naive about digital transformation in healthcare than we were previously. I think I have benefited from the experience of the last 15 years, and I think it's made me a little bit less likely to to get it wrong. But think about the beginning of the social media era. Like, who could have predicted that it would completely disrupt our politics or, you know, or the impact on teenage mental health, etc., etc. these things are, I think, inherently unpredictable, fluid. Have second, third, 12th order effects. So I think pretty hard to predict. That said, we're in the predicament that we're in. This is rolling out. And as I sat down and spent a year and a half of my life trying to think really hard about what do we know? I spoke to about 105 people from all different walks of life, from the tech industry, from healthcare, from government, etc., and did my best to, you know, and armed in part with the wisdom of having screwed up the last time and, and underrated, underappreciated, the unanticipated consequences, at least my best guess is for the next ten years or so.
Dr. Robert Wachter: [00:31:49] These tools have the capabilities of solving a whole lot of the problems that are inherent in the current system And those problems include all the administrative paperwork, all of the friction that patients see when they try to see a doctor. The fact that I, as a generalist, when I'm on the wards, you know, I'm pretty decent hospitalist on a good day, but I'm a generalist. And every patient I see, there's somebody in the building who knows more about their specific problem than I do. And in the old days, meaning three years ago, the only way I could take advantage of subspecialty level knowledge if I needed to curbside an oncologist, I would hope to run into you in the cafeteria. And now I can pull out of my pocket a tool and say, I have an 82 year old patient with Waldenstrom's who comes in with a fever, a white count of 12, an infiltrate on chest X-ray, an abnormal lfts, and a creatinine of three. What do you think is going on? You know, if I put that into up to date or Google, it would have choked.
Dr. Robert Wachter: [00:32:46] I was like, what are you talking about? I can now get subspecialty level knowledge anywhere and so can in some ways, so can patients. So in some ways the system seems quite well-designed in some ways to solve a lot of the problems. And I think that part of the golden Age is if you take medicine as it currently sits and say, if you could remove a lot of the administrative barriers, the bureaucratic, the paperwork, the stuff that is silly and adds no value to patients or to doctors or nurses, and could scale the knowledge of subspecialists and of the medical literature to make it more readily available to everyone. And the electronic health record is no longer just a filing cabinet. It actually is a source of insight and intelligence and suggestions. Diagnosis or the right tests. Then you liberate the system from a huge number of the flaws that it has, and then you're left with a system where if you take a step back and get rid of all the craziness, you say, we can now treat obesity, we can now treat many cancers that we couldn't treat ten years ago. I mean, the system itself should be in a golden age. And so part of my enthusiasm is your. I think this the capacity of the new AI to remove a lot of the stuff that in some ways prevents this from being a Golden age.
Dr. Robert Wachter: [00:34:00] That's the opportunity that I see here. Are there things that could go off the rails? Absolutely. As we move, I like the way you're framing the sort of the H and P and the and the data collection as one set of skills, and then another is the sort of assessment and plan and the recommendations and what do we do? And absolutely, as the AI begins intervening in that world and suggesting and maybe in some ways prescribing, there's a lot of mischief that can happen. The AI is now suggesting the right treatment. You can imagine a world where there are two treatments. One is 10,000 bucks and one is 500 bucks. And in some ways, perfect world. Of the $500, one is as good as the 10,000. And that's what it recommends. Great. But maybe in a world where the health system gets paid more for the 10,000, that's what it recommends. Or there's some conflict of interest that is influencing the recommendation. So ask the AI gets more prescriptive and more autonomous. There's some funky stuff that can happen. That may be really problematic. I'd say the biggest problem I start off the book with the idea of a digital twin, which is the Mayo Clinic is experimenting with having a trusted digital twin of a Mayo doctor, being able to scale Mayo care to more people in more places. That's pretty exciting.
Tyler Johnson: [00:35:12] For listeners, this is like a pseudo embodied avatar.
Dr. Robert Wachter: [00:35:16] It is a doppelganger of the doctor, but trained not on the entire internet, including The Onion and Reddit, but trained on Mayo or Stanford or UCSF data so that, at least in theory, it would say to a patient exactly what I would say. And the CEO of the Mayo Clinic said to me, you know, 20 or 30 years ago when we wanted to deliver Mayo care to more people in more places, what do we do? We built a campus in Scottsdale. We built a campus in Jacksonville. We would not do that today. We would say, how do we use AI to scale Mayo expertise and Mayo's brand to more people in more places? That's very exciting. On the other hand, that's got a massively dark side, which is you could take an audio of what I'm saying to you now and have me and my voice say, you know, you shouldn't get vaccinated. Those things will kill you. So the dark side of a digital twin, a trusted digital twin, is a deep fake that can basically say anything. So there's a lot of stuff that can go wrong here. But part of my enthusiasm is that I actually also don't think that doctors or nurses jobs are going to be compromised for the foreseeable future. I think that the unmet needs are so great that if these tools relieve us of paperwork and allow us to be more efficient in our work and maybe, hopefully happier in our work, I still think we need all the doctors and all the nurses that we have now.
Dr. Robert Wachter: [00:36:30] I do think there are a decent number of job losses, but they're in the billing department, they're in the call center. There are if you look at the job growth in healthcare, and we have been a relentless job creator in the US over the last 20 years, it's been far more in administrative jobs than it's been in clinical jobs. So that's also part of what gives me hope, in part politically because of doctors and nurses thought their jobs were at risk. They would push back pretty hard on this, and they wouldn't say, I'm worried about my job. They'd say, this thing is going to kill you. And so I think all the forces are aligned to have at least the next ten, 15 years be net pretty, pretty positive. Beyond that, who knows? I mean, I don't think I am smart enough or clever enough to think about what does this look like 15 or 20 years from now? It could be awful. And again, everything I'm talking about. By the way, is all healthcare. I'm very worried about AI in the rest of our lives. I mean, if there's 20% unemployment because of AI or people who are developing bioweapons, etc., you know, that's not bad. I'm talking about healthcare and healthcare. Part of my enthusiasm is the healthcare system is in such dire straits. I think we need the help.
Tyler Johnson: [00:37:31] So I want to parse apart two different sides of the argument that I hear you make. So one thing certainly I think it is true. I think it's maybe inarguably true that there are problems in the current health care system, that there are certainly use cases for AI being able to help. Right. I think one, reading your story about getting this kind of. And you make this point in the New York Times article in the book about getting a sort of an Insta consult, a consult in your pocket. Right. That made me think of a hospitalist, actually. So thank you for that here, who I have worked with for a number of years. There was a time when his team of Stanford residents was receiving a patient who had been cared for in an outside hospital, and they're kind of sitting around bantering back and forth in the workroom, and the attending is sort of over at the side. They're not rounding. They're just kind of having conversation, not really realizing that he's listening, and he hears them saying some unkind things about the care that was provided to the patient at the community hospital from which the patient came. And it turned out that this particular hospitalist had spent a couple of years working at a hospital.
Tyler Johnson: [00:38:40] And, you know, attendings don't always do this, but he kind of put his hand up, called for silence and said, hey guys, do you realize here at Stanford, if you have a question about nephrology, you can call Glenn Chertow or whatever other nephrologist and all they will do is help you with the kidneys. Yep. And if you have a if you have a question about infectious disease, not only can you call the infectious disease service, you can call the immunocompromised solid tumor, infectious disease or whatever, right? There are specialists within specialists within specialists here who can give you the most expert advanced advice, according to the most cutting edge research and the most vast store of experience and knowledge, maybe in the world in some cases, right? If you're at a community hospital, you may be in a place where you don't have a dialysis machine, let alone a nephrologist. You may be in a place where you can hardly get, you know, an advanced read on a blood smear, let alone consult your friendly neighborhood hematologist, as you said. When you run into in the cafeteria.
Dr. Robert Wachter: [00:39:41] Absolutely. Yep.
Tyler Johnson: [00:39:42] So having open AI or whichever, you know, company having a consult in your pocket so that a person who's at a community hospital who just doesn't know things because they have no reason to know them, because they've never done fellowship in 27 different internal medicine, subspecialties, whatever. Certainly that seems like an inarguably good thing, right?
Dr. Robert Wachter: [00:40:02] Assuming it gives you correct and trustworthy advice, which, which it does. And it's and they're only getting better.
Tyler Johnson: [00:40:09] Okay. So that I will stipulate, but I also want to push back in a way that to me is the thing that worries me the most, which you've kind of alluded to, but I would like to hear your answer to the the real sort of substance of this. Let me frame my concern by talking about two different use cases. One you mentioned very briefly is social media. I continue to believe that social media has the theoretical potential to revolutionize the world in good ways. I think that the technology of social media, if it were employed in a way that where its bottom line, so to speak, was human flourishing and community building and the propagation of correct information and whatever, I think it could do incredibly beautiful and powerful things.
Dr. Robert Wachter: [00:40:54] Well, not not we know that because that's what Twitter was during Covid.
Tyler Johnson: [00:40:57] And I remember right when I was an attending here. I mean, I'm an oncologist, but during Covid, everybody's read everything, right? When I was an attending here and we started to get news in the spring of 2020 about what was going on in Italy, immediately one of my Co-residents chiefs residents from residency formed a Facebook group. We had doctors from across the country, and we were in real time sharing best practices and, you know, all sorts of things in a way that would have been impossible even ten years before. And that's a micro example. That was just, you know, our people that we knew, but even that was, as you say, a powerful testament to the possibility. And yet what has actually happened is that you can read about the implosion of Twitter under the leadership of Elon Musk. You can read about Facebook and it is a matter of congressional record now, right? That Facebook knew as just one example that they were harming the mental health of teenage girls in particular, and yet prioritized profit over the health of teenage girls. You can read about both Instagram and Facebook. There is a long track record of articles in the Times and Wall Street Journal of times when they have, for example, propagated the sexualization of young women, purposefully feeding those pictures into like a, you know, a ten year old girl in her bathing suit, purposefully propagating those through algorithms to luring older men because it makes them money.
Tyler Johnson: [00:42:21] Right? So that's one set of examples over here. Then on the other hand, if we look at the EHR, in my mind, to be clear, what has happened with the EHR as opposed to what we hoped would happen with the EHR, is very complicated and has a lot of causes, but certainly one of the causes is that healthcare companies, not doctors and nurses, but healthcare companies, figured out that the EHR was an absolute goldmine for ringing higher profits from healthcare because, as you mentioned, you could look at a healthcare record and it's not about being dishonest, but it's just about, for example, trying to make sure that you are accounting for the actual complexity of a patient so that you can get paid accordingly becomes a task that then largely, at least in the first iteration, this is changing, as you've discussed to some degree, but largely fell on clinicians right. And there. And like, I know you're in Stanford, for example, but there was a huge corporate push here to standardize admission records for patients admitted to the hospital, precisely as I was just talking about, to ensure that they fully reflected the complexity of the patients. And on the one hand, you can say, well, sure, having accurate medical records is a good thing.
Tyler Johnson: [00:43:36] Right. And I'm not arguing about that, but it's all to say that the push behind that, it wasn't that it improved care. It didn't change the care at all. But what it did do was it improved revenue. And so the point that I'm trying to make is that the common thread that I see between both of those things is that in both cases, in effect, what happened, in my view, was that a technology that had the potential to do powerful things for good. In one case, in social media, I would say it has largely had negative effects. Not to say it hasn't had some positive effects, as you mentioned, but I think net, net, it has been overwhelmingly negative. In the case of HR, I think it is clearly not some medical things as we discussed earlier, but I think it has also had many more negative effects that are largely a reflection of the profit motive, rather than their reflection of the problem. That profit has been prioritized over the welfare of clinicians and patients. And so the thing that I worry about is that for all of the magically amazing things that AI can do. I don't see it solving that. I don't see it removing profit as being the driving force behind most healthcare corporations, which largely, you know, it used to be doctors would kind of hang out their shingle and practice independently.
Tyler Johnson: [00:44:48] But largely now most doctors are employees rather than entrepreneurs. And so I fear that almost no matter how AI transforms healthcare, that eventually it will in effect be co-opted. For example, if it makes people more efficient, then healthcare companies will simply see that and give doctors more work to do because now they're more efficient than they have more time. So they should be given more to do, right? And then on top of that, as you mentioned, is that if AI bots are controlling diagnostic and therapeutic algorithms, especially autonomously, but even if they were even with a physician overseer, if you then introduce even just a, you know, it's like putting a drop of red dye into the pool, even if you introduce a little bit of a profit motive, right? A pharmaceutical company that backs the development of an AI bot, and then the AI bot just gives a hair's more preference to this beta blocker than that beta blocker or this chemo regimen, than that chemo regimen. That seems like a, to your phrase, from before a foreseeable potential case that could be at least detrimental and potentially catastrophic. And I guess I just maybe I'm feeling a little pessimistic watching social media for HR and everything else. And also the developments with Sam Altman and OpenAI and whatever, why not subscribe to that brand of pessimism?
Dr. Robert Wachter: [00:46:05] That's fair. I mean, I there's nothing about AI that transforms the dominant incentives in the system, which are mostly kind of screwed up, and that decreases the risk of the kind of financialization of healthcare and, and in some ways, profit maximization. I guess the one thing I would disagree with you about is you paint the doctors as inocencia as being the victims of all these corporate overlords, whether it's UCSF or Stanford Healthcare or UnitedHealthCare or epic or whatever, we're part of the system. And although, you know, we're unhappy about having to click extra boxes for better billing, on the other hand, we're in most systems paid based on the coding and therefore, and probably the committee that decided to do whatever it was that it did included physicians who said, you know, I mean, the reason we all sleep at night is the revenue in a place like yours or mine is partly to make sure that we can continue to educate trainees and deliver care to patients with no insurance. And and, you know, and do you know, world class research that needs subsidization. So I don't think there's any innocence in this to the extent that everybody has pure motives and, and is not potentially going to be influenced by the financial system and absolutely right. I think you can wonder and worry, and I think it's reasonable to do so to say here is an increasingly powerful technology that will be potentially under the control, maybe in ways that are not completely transparent by some very big companies that have access to the data, that then are giving recommendations and maybe increasing number of cases, acting autonomously without someone who took the Hippocratic oath looking over its shoulder.
Dr. Robert Wachter: [00:47:46] And that is influencing the care that is being delivered and recommended. And it would be naive to say that everybody in that system, I'd say, including the docs, is to some extent going to be profit maximizing. And so the more powerful it gets, the more agentic it gets and the more it is making the decisions. I think that is something we should absolutely worry about, probably something that requires some level of regulatory engagement. I think if we were a more functional political system today, we would have a really high level group saying, how do we regulate this? How do we ensure that in the balance between the motives of. The Hippocratic Oath and the motives of Gordon Gekko, and profit maximization that we tilt in the right direction. But we don't have that system, and we have a system where the market is kind of allowed to do what it does. So I think your your concerns are appropriate. I'm worried about them too, but I guess I'm just left with the don't compare me to the Almighty. Compare me to the alternative. The alternative is a system where there's already that kind of stuff going on and a huge amount of waste, bureaucratic silliness, time that doctors are spending doing stuff that adds no value to patients. This issue of if, let's say, AI scribes, which is really the first ubiquitous use case of the new AI, saves me two minutes per visit. Do my corporate overlords say, great, I'm going to make the hamster wheel spin two minutes faster and have you see an extra patient every session? Maybe.
Dr. Robert Wachter: [00:49:17] But part of the reason they are paying 2 or $3 million to nuance or a bridge or whoever your company is that does this is physician burnout and is patients saying, well, when I went to see my doctor at Stanford, I noticed that he wasn't looking me in the eye. He was looking down at his computer. Your system has an interest, including a financial interest, in keeping you happy. Happy ish? Well, maybe not ecstatic. Yeah. It has an interest in having the patients believe that your doctors are engaged and empathic. And that gives them a motive. So that I'd say if they save two minutes of time, I think that that a rational system, including one that's paying a lot of attention to the bottom line, will take one minute of that time and make the hamster wheel spin a little faster and give you back one minute of that time. And I'm not being naive about this. It costs about $1 million to recruit and retain or replace a physician. So it is not in the corporate interest of the health system to say, oh, you know, physicians were complaining. The burnout rates were incredibly high. Pajama time. We're going to spend a few million dollars to bring in AI scribes. And the minute we do that, we're just going to make the hamster wheel spin that much faster. So you're going to be working just as hard as you did before. There's actually is a corporate financial interest to make your life a little bit happier.
Tyler Johnson: [00:50:27] Yeah. I don't mean to be totally I'm actually largely an optimistic guy though, about tech. I think I'm more pessimistic than a lot of people around here. I guess I just I worry in some ways that what we are coming up with are more efficient ways to do lots of things, which is great unless the thing you're doing is actually a bad thing, and then our efficiency just gets you to a bad place faster. Yeah. And I do worry that when I look at the actions of people, like a lot of the CEOs and other corporate officers have a lot of the most powerful social media and general tech companies. I wish they inspired me more confidence that their prime directive was more about human flourishing and not about personal profit. So in the conclusion of your book, your final section is called something like Being a Physician in the age of AI. And so you've talked about all of these, you know, thousand different use cases and whatever. And then you kind of bring it right back to your own experience. And you talk about the experience of attending on the UCS, UCSF wards as a hospitalist, attending around the time you were finishing your management. So then you go through and you say, well, you know, here I was operating as a doctor and, you know, I was using AI in my pocket for these things. But it was clear to me, even as I used it, that it couldn't do all of these other things. And then I thought about, well, what have you been in effect? What if you could do all of these things? And then if I had this sort of magical AI bot that could do all of the things, do I think I'm Bob Wachter, think that my patients would want to have me replaced by whatever the bot would be? Your sort of final flourish is saying, I think in effect at the end of the day, that there is always going to be some irreplaceable part of doctoring where people are going to want not a robot, but a doctor.
Tyler Johnson: [00:52:15] So my question is, What gives you that confidence? What if a bot became better at diagnosing and coming up with treatment plans? I mean, to your point, the humans are not your doctors are not angels. Doctors are certainly not perfect diagnosticians. They're certainly not perfect coming up with treatment plans. But if a bot was better at all those things, what if a bot was better? You talk about multiple times in the book about using a bot to come up with the way you're going to have a hard conversation with the patient. What if what if a bot was better at having a hard conversation? It seems paradoxical to me that on the one hand, we're talking about all the potential problems with AI, but then let's say that all those problems get solved and the bots actually are better. Why are you confident that patients would still want a fleshy human being as their doctor rather than a bot? If the bots really got that.
Dr. Robert Wachter: [00:53:01] I guess the short answer is I'm not super confident about that. And I can imagine a world where the bots are so good that patients can choose to see a human doctor, but have to pay the full incremental cost of what it would cost to see the human versus the bot. I mean, when I have it, when I pull up my phone and a Waymo is ten minutes away and an Uber is two minutes away, I will wait for the Waymo. I prefer being in this car with no driver, so I can either take a nap or sing Springsteen at the top of my lungs. My taxes are complicated enough that I still see a human accountant, but if there were a little less complicated or I would be perfectly comfortable using TurboTax. Same thing with cruise with travel agents. So I think this is an open question. I can see a world where absolutely, the AI is good enough that for many people it is more convenient, it's less expensive, it's available at their fingertips that they will choose to get it. In writing that last chapter, I really tried to fantasize and this was really specific to my role as a hospitalist. So here I am in a big, complicated teaching hospital and patients who get hospitalized. It's not like the old days where they came in because they had community acquired pneumonia, or they had a small PE. They they've got nine things wrong.
Dr. Robert Wachter: [00:54:11] And in that world of complexity, I could not see in the foreseeable future that the AI, even if it's really good at making a diagnosis and making a treatment recommendation, can handle all of the complexity, the coordination, the fluidity of of care, the operational challenges, at least in the foreseeable future. Maybe that's over a ten year time horizon. I'm also a little bit influenced by what is the most replaceable field in medicine. Without question, it's radiology. The founding father of generative AI, Geoff Hinton, in 2016 predicted that we wouldn't need any new radiologists in 2021, and by 2021, we should stop training radiologists. And he was clearly wrong at UCSF and I'm guessing at Stanford, we can't hire radiologists fast enough in an era of this unbelievable AI. So I think predictions of the demise of physicians are overstated. I do think what we will start looking at is task specific. And to say, for example, in primary care, what is it that a primary care doctor does that really is extraordinarily human, at least we think. The way we think about it today involves a level of complexity, often social and economic complexity, and what does not. And to me, there's a study a few years ago said the average primary care doctor, if all they did was prevention, that takes 27 hours a day. If no patient had the temerity to actually be sick, your job is 27 hours a day. It's impossible.
Dr. Robert Wachter: [00:55:35] And so can you. Imagine a world where we say your blood pressure, your wegovy, your cholesterol, your vaccines, you know, your travel medicines are all managed by AI, maybe with some escalation strategy. When you fall off the algorithm, it goes to a doc. Yeah, easy. I think that's the way this goes where we look at the tasks, including for specialists like an oncologist and say, what is it that you are doing that really takes your level of training and deserves your income and says, that's half of what you do, 80% of what you do? I don't know, but there's some piece of what you do that right now. We say, oh, it doesn't need a physician. Let's have a PA do it. Or let's have a nurse practitioner do it. That's a decision we made 20 years ago to say it doesn't need a doctor. It needs someone who's lesser trained and less expensive. We can take the work and parse it. I think we're going to do the same thing with AI, but I still think there's going to be plenty of work for doctors. And at the end of the day, I do think that most humans, when they learn they have cancer, are going to want to hear it from a human and have a conversation with a human. Am I confident that's going to be true 20 years from now? No, I think that's a that's a pretty long time horizon.
Tyler Johnson: [00:56:40] Let me ask one final philosophical question. Just imagine a world where, you know, some future iteration of a bot really can 1 to 1 match what a human can do? All the complexity, all the nuance, all the dynamic changes in physiology, probation, the whole nine yards, they can even have a conversation. They can pass the Turing test, the whole for the most complicated cancer, end of life, whatever, whatever, whatever. The whole thing in that scenario. Do you still think there is some kind of intrinsic value add to interfacing with the human as opposed to a robot?
Dr. Robert Wachter: [00:57:21] I think it's completely unpredictable. I think that the people will determine that. I think the thing that is a little hard to factor in, in that is in some ways the politics and the politics would be, if that happens, that for medicine, the most complex human of fields, then it's already happened in law, in accounting, in journalism and consulting, it means the unemployment rate is 70%, and it means we have a revolution because that will be I don't believe that you can solve that problem with universal basic income. And I'm sitting at home playing chess every day and collecting my check. I think humans get real value out of working. I think the societal upheaval that results from that kind of job replacement is going to be massive and makes any prediction about what that looks like completely essentially unknowable.
Tyler Johnson: [00:58:12] Thank you so much for being here and thanks for joining.
Dr. Robert Wachter: [00:58:14] It was a joy. Thanks so much for having me.
Henry Bair: [00:58:18] Thank you for joining our conversation on this week's episode of The Doctor's Art. You can find program notes and transcripts of all episodes at The Doctor's Art.com. If you enjoyed the episode, please subscribe, rate and review our show available for free on Spotify, Apple Podcasts, or wherever you get your podcasts.
Tyler Johnson: [00:58:37] We also encourage you to share the podcast with any friends or colleagues who you think might enjoy the program. And if you know of a doctor, patient, or anyone working in healthcare who would love to explore meaning in medicine with us on the show, feel free to leave a suggestion in the comments.
Henry Bair: [00:58:51] I'm Henry Bair.
Tyler Johnson: [00:58:52] And I'm Tyler Johnson. We hope you can join us next time. Until then, be well.
You Might Also Like
LINKS
Explore Dr. Wachter’s books, articles, Substack, and more.
Read about the history and evolution of hospitalists.
