The Lost Aura with John Lantos, MD

Announcer

At The Nocturnists, we are careful to ensure that all stories comply with healthcare privacy laws. Details may have been changed to ensure patient confidentiality. All views expressed are those of the person speaking, and not their employer.

Emily Silverman

This is The Nocturnists. I'm Emily Silverman. Today's episode features Dr. John Lantos, pediatrician, bioethicist, and physician-writer, whose latest essay, The Lost Aura of the Physician in the Age of Artificial Intelligence, asks a deceptively simple question. What happens to the role of the doctor when artificial intelligence becomes remarkably good at practicing medicine? John argues that AI didn't create the challenges facing modern medicine. It simply accelerated trends that have already been unfolding for centuries. As medicine has become increasingly data-driven, standardized, and technological, the relationship between doctor and patient has gradually shifted. Now, with large language models offering medical advice, empathy, and even shared decision-making, we're forced to ask what remains uniquely human about the work of healing.

Emily Silverman

In our conversation, we explore the history of the doctor-patient relationship, the concept of the physician aura, and the surprising possibility that AI won't replace physicians so much as redefine what they're for. But before we dive in, take a listen to John reading from his essay, The Lost Aura of the Physician in the Age of Artificial Intelligence.

John Lantos

French philosopher Michel Foucault analyzed a shift in the way doctors interacted with their patients in late 18th century France. He showed how in the minds of these scientific pioneers, individual physicians and patients were increasingly framed as what he called disturbances that can hardly be avoided, but that ideally should be neutralized in the pursuit of objective medical truth. Foucault called this abstraction the clinical gaze. The patient's story and the physician's interpretive presence became secondary to precise measurements and visualized pathology, quantification, and standardization. Physicians were encouraged to spend less time talking to patients and more time gathering the data that would lead to an evidence-based diagnosis and treatment plan. AI is the logical outgrowth of this philosophy of medicine. It can perform many of the non-technical tasks that were once thought beyond the reach of technology and machine learning. It now obtains informed consent, counsels patients about end-of-life decisions, and provides psychotherapy. It does so with language that expresses attentiveness and concern.

John Lantos

Perceived compassion and patient explanation no longer depend solely on the capacity of clinicians. These capacities of AI might allow it to fulfill the dreams of both biotechnological medicine and humanistic medicine.

Emily Silverman

I am sitting here with Dr. John Lantos. John, thank you so much for coming on the show.

John Lantos

It's great to be here. Thanks for having me.

Emily Silverman

So I want to get into your piece, The Lost Aura of the Physician in the Age of Artificial Intelligence, because I think it's a really interesting topic. But before we get there, I would love to learn a little bit more about you. You have a long and storied career in pediatrics and in bioethics. So I was wondering if you could maybe paint for us a sweeping arc of how you came to be a pediatrician and also how you got involved in bioethics.

John Lantos

Sure. I was always interested in humanistic aspects of medicine. I was an English major at Brown before going to med school. And when I got to med school, became fascinated with the ways that medicine was changing. One of the experiences I had that shaped my view was after my second year of med school, got to go to Nigeria for three months and spend some time in one of the teaching hospitals in Benin City, Nigeria, and came to realize how quirky Western medicine is and how extravagant it is. I saw patients with neonatal tetanus, measles, patients who were dying of dehydration, and compared those with the patients I was seeing as a medical student in Pittsburgh, children with leukemia or tiny premature babies in the neonatal intensive care unit, and realized just how much effort and energy we put into saving lives when many places around the world, lives can be saved with a little oral rehydration solution or better access to vaccines or other inexpensive preventive or therapeutic medicines.

John Lantos

I did my pediatrics residency in Washington, D.C., Children's National Medical Center, and those were the early days of neonatal intensive care. I was a resident in the early 1980s, and that was before neonatology fellows were in the NICU at night. So as second-year residents, we were making life and death decisions for babies without much evidence. Most of what we were doing was exuberant innovation with no rigorous studies or long-term follow-up to see whether, A, it was working acutely or led to long-term complications. During those years, there was also a national controversy about neonatal intensive care, particularly for babies with Down syndrome. There was a case called Baby Doe, which involved parents refusing life-sustaining treatment surgery for their baby with Down syndrome and esophageal atresia. And this got me hooked on the ethical aspects of neonatal intensive care. After residency, I spent a couple years in the National Health Service doing primary care in southern West Virginia, which also gave me insights similar to working in Nigeria, where I saw the effect of health disparities, in this case, for rural Americans rather than Nigerians.

John Lantos

The differences were striking. Most of the technologies that had been available in the tertiary care center where I trained and that I had learned to use well were not available. So I had to spend more time talking to patients and cobbling together treatment plans with whatever resources they had access to. I wanted to pursue some of the ethics questions and was lucky enough to find a fellowship at the University of Chicago, one of the first fellowships in medical ethics, and went there to do the fellowship and ended up staying for 20 years at the University of Chicago. I had some brilliant mentors. Mark Siegler, who started the MacLean Center for Clinical Medical Ethics, one of the first academic medical ethics centers, doctors and a pediatrician named Arthur Kohrman, who had been head of the American Academy of Pediatrics Committee on Bioethics and was running a specialty hospital for children with complex chronic disease.

John Lantos

So he hired me as a clinician. I worked there clinically and worked in the Bioethics Center and had a wonderful 20 years with some amazing colleagues exploring all the issues around innovative therapy like neonatal intensive care, liver transplantation, bone marrow transplantation, and also some of the policy changes and changes in the way we were delivering healthcare. I had the opportunity to participate in the health reform task force that President Clinton started in the early 1990s and to look at the ways that things like HMOs and other finance arrangements were changing the doctor-patient relationship. So we did a lot of writing about that and about the dilemmas of research and innovative therapy.

Emily Silverman

I'm wondering if after this rich 20-plus year career of clinical medicine and bioethics, if there are any like macro takeaways that you have, did you come away feeling more sure or less sure about ethical frameworks? Are there any ways of thinking about ethics that solidified with you over those 20 years? And I'm just wondering like coming out the other end of that, how that influenced the way that you think about bioethics and medicine and the world in general?

John Lantos

My thoughts on those big topics evolved steadily. I mean, I ended that story in Chicago around 2007. After that, I went to Kansas City and started a pediatric bioethics center at Children's Mercy Hospital. Through it all, realized that there was this synergistic overlap between what was happening in bioethics, what was happening in health policy, and what was happening through advances in medical technology, that were all working to strengthen and solidify a reductionist approach to thinking about health and disease that was enormously powerful. It was the idea in the quote that I read from Michel Foucault at the top of this, that the more objective we can be, the more we can get measurable data and molecular understandings of what causes disease, the more we can develop targeted treatments, personalized interventions, and ideally cure diseases that resisted medical intervention since the beginning of time.

John Lantos

At the same time, these were extraordinarily expensive interventions, so unlikely to be available to most people in the world, and they were interventions that led people to move closer and closer to that ideal that Foucault described, where the individual doctor and the individual patient would become irrelevant because the objective data about disease and the evidence-based guidelines for treatment would make it so that we didn't need to listen to the patient, hear their story, understand what illness meant to them, and we didn't need to form a trusting, deep physician-patient relationship because, as we learn in academic medicine, doctors switch off with one another frequently. One doctor will be on this week, the end of the week they leave, another doctor comes on, and we act as if nothing is lost when that happens because we act as if the trusting relationship between an individual doctor and an individual patient is either irrelevant or of minimal importance compared to the objective data.

John Lantos

I think this leads to dissatisfaction on both sides of the doctor-patient relationship. I think doctors find it morally unsatisfying to work in this way. The relationships that make medicine a rewarding profession are devalued, and patients find it impersonal, dehumanizing, and unsatisfying for similar reasons. So there's sort of a paradox. We built this system, we spend now $5 trillion a year, an absurd amount of money, I mean, more than the GNP of all but three countries in the world, on healthcare, and have built a system that gives us worse health outcomes than many other countries at a much higher price with higher levels of dissatisfaction. It's quite a paradox.

Emily Silverman

And then AI comes into the picture, which is the topic of this essay. When did you first realize that AI had made this quantum leap in ability and that it was going to start mixing in with clinical encounters, and have you had a personal journey around your opinions with that? I mean, I think this piece obviously presents, I don't even want to say it presents an opinion, it's really more of a question or a place to start a conversation from, but I'm just curious where you've gone with the AI conversation.

John Lantos

Yeah, I remember, what was it, 2023 when ChatGPT-4 was first released, and I started playing around with it, and was fascinated, impressed, and frightened by the potential of this. That stimulated me to really start delving into what people had been writing about artificial intelligence, a field that I really hadn't been paying much attention to before that. And I discovered that lots of people were way ahead of me on that interest, on that fascination. People like Eric Topol, who wrote a book called Deep Medicine five years before ChatGPT-4 came out, and was already aware of some of the opportunities and some of the challenges.

John Lantos

Robert Wachter wrote a similar book. Zak Kohane and colleagues had been asked to evaluate the potential of ChatGPT before it was even released for their practice of medicine. And they wrote a book about this. Sherry Turkle, who was a psychologist at MIT studying human interactions with early AI models, who started reading all of that stuff and then playing around with the effectiveness of ChatGPT and eventually other LLMs, large language models, both for my practice as a general pediatrician, but also because I was starting to have more and more health problems for my knowledge and empowerment as a patient. And so from both sides, both the doctor side and the patient side, started to see the enormous power that these large language models had for changing the way we practice medicine and changing the way we experience healthcare as patients.

Emily Silverman

So the title of the piece is The Lost Aura of the Physician in the Age of AI. And you talk about this aura concept, which you borrow from philosopher Walter Benjamin. Explain aura. What is that? What does that mean? I think, you know, probably there are different meanings in the world of psychics and things like that. But Walter Benjamin uses the word aura to refer to something that's a little different. So walk us through that.

John Lantos

So Benjamin, he was a German Jewish philosopher writing in the 1920s and the 1930s. And one of his most widely cited influential essays was not about medicine or AI, but it was called The Work of Art in the Age of Mechanical Reproduction. And Benjamin was watching what was happening in film at the time, particularly the way the Nazis were using film to engage the masses in the fascist project that the Nazi party and Adolf Hitler were using to build their nefarious regime. And he saw the way film could manipulate the masses, could be used both as a tool as well as an art form. It led him to question what we mean by art in an age when this new technology was coming along. So he looked at film, he looked at photography and asked, what's going to happen to painting? What's going to happen to the idea of an original work of art in an era where visual art can now reproduce any image and transform it so that we can see things we never saw before seems to be a more powerful technology for conveying images than had ever existed before.

John Lantos

And yet there are these things called paintings where people sit for a portrait and the painter labors over the portrait for weeks or months or sits in the garden watching the light change on the water lilies. And he came up with the idea that original works of art have this thing that he called an aura, which he never precisely defined, but what he meant by it was something unique about the circumstances in which it was created, the time and place in which this particular moment was captured and the particular skills that were necessary to do that. In the essay, I take that idea and make an analogy to what's happening in medicine now. I suggest that something like Benjamin's concept of aura existed since the beginning of time in relationships between doctors and patients. The clinical encounter happened at a particular time, a particular place with a particular conversation between two unique human beings.

John Lantos

And that that was something people valued the way they valued original oil paintings, but something that in the age of AI was transformed, whether it was uniquely valuable or irrelevant. That's the question that I try to get at in this essay, because it seems to me what happened to painting in the late 19th and early 20th century is happening to medicine in the late 20th and early 21st century. That unique physician-patient encounter is now seen as almost an anachronism.

Emily Silverman

And you talk about how this AI technology shifting the way that we think about or that we value an individual doctor-patient interaction in a particular time or place, that this is almost the climax of a process that's already been unfolding. So you talk about Foucault and the clinical gaze, which you already mentioned, but you also talk about other changes that have happened in medicine, like anesthesia, turning patients into passive bodies, evidence-based medicine. Obviously, we want to be making decisions based on evidence. But like you said, sometimes you can lose the importance of that trusting relationship. You also talk about the electronic health record. So AI is not really the beginning of this change. In some ways, it's the crystallization of something that's been happening maybe for a long time. I don't know if you want to expand on that a bit.

John Lantos

Sure. So just the way Benjamin recognized that film and photography could democratize art and make art more accessible to the masses. I think AI arrived at a moment when medicine was already far along in the transition that Foucault described as beginning 200 years ago. I talk about a bunch of things. Some of them are the ones you mentioned. But another big one was just the rise of the hospital as a locus of medical technology. So if Paul Starr has shown in his book, the social transformation of American medicine, that a shift occurred probably early 20th century from medicine as an art that was delivered in the home. Doctors made house calls. When they made house calls, they saw not just the patient, but the patient's milieu, what their home looked like, what their family looked like. And the patient, it was much easier to see the patient as a unique individual. Once they go to the hospital, all that is lost.

John Lantos

And the transition from home care to hospital care was an important step along this trajectory by which patients lost their individuality or doctors stopped perceiving individuality as important compared to the more objective measures that we could do when we had x-rays and lab tests and all the things that are gathered together in a modern hospital. So anesthesia and surgery were part of that too, because they were delivered in the hospitals where patients had lost their identity by going into the hospital and then were rendered completely unconscious so that the work of the surgeons and the anesthesiologists could be done in a safe and effective way. The individuality of the patient was less relevant than the safety of the physiologic monitors and technologies that allowed surgery to be done safely and effectively. Evidence-based medicine came along and accelerated those trends. It said, well, if you think about the randomized control trial, which is the gold standard for evidence-based medicine, the fundamental assumption of that is that patients are interchangeable.

John Lantos

We don't want to know anything about their individuality. In fact, the power of the technology is to erase all traces of individuality because they're seen as biases or confounders. They don't help us understand what's effective medicine. They get in the way of that understanding. This again led to ways of measuring the effectiveness of medicine that didn't rely on subjective reports but relied on objective measures of outcomes that reliably reproduced. So all of those trends set the stage for AI. AI fit into them perfectly, but it didn't create them.

Emily Silverman

So do you see AI as an answer to the problem of the dehumanization of the patient in such a way where we're actually able to rehumanize in some way, even though it's not an actual person or consciousness, like I'll say, and I'd love to hear your experience too as a patient, like I do get a sense of comfort and relief from just being able to pop a question into ChatGPT and I'll put in my blood tests and ask it to give me its thoughts. And it does so immediately. And as a physician myself who's gone through medical school and residency, it's usually pretty good.

John Lantos

Right. And getting better all the time.

Emily Silverman

And getting better all the time. And it is comforting and it is infinitely patient. It's definitely empathetic, sometimes to the point of being sycophantic. But it listens. And I don't know, I'm just, I'm curious what your experience is with AI. And then also, what do we know about studies of how patients perceive AI as a provider? And does it bring them comfort? Do they trust it more? Do they trust it less? There's accessibility differences.

John Lantos

Yes. I'll tell you a story about some experiences I've had as a patient. I've recently had some heart problems and had a Holter monitor done where they did that two weeks of electrocardiogram tracings. And the Holter came back showing that I had some life-threatening cardiac arrhythmias. I was having runs of ventricular tachycardia. I was asymptomatic. I'd had a prior heart attack. So people thought this was related to the prior heart attack. But because I was asymptomatic and generally was doing well, my doctors were uncertain what the next step was. So some people talked about putting in an implantable defibrillator. Some people talked about a pacemaker. Some people talked about anti-arrhythmic medication. And one of my favorite responses from a doctor was a guy who said, look, you are in an evidence-free zone. We have no idea what to do with a patient like you. That was not particularly comforting. So I turned to the LLMs, gave them all my information, put in the Holter tracing, put in my labs, put in cardiograms, even my cardiac MRI, and got very valuable information about the reasons why there was no good evidence about the best way to treat me.

John Lantos

All of the LLMs seem to be programmed to say, of course, we can't give medical advice before they then give you medical advice. They always say, consult your doctor, but they give some information. Here is one of the most interesting things I found. We know that when doctors talk to patients about treatment options, when they engage in the process of shared decision-making and seeking informed consent, doctors bring their own biases and their own practice styles to these discussions. So I was uniquely empowered as a patient because I knew lots of cardiologists personally. So I could call up one, get one bit of advice, call up another, get another bit of advice. Most people can't do that. But with the large language models, you can create the kind of doctor you want. So I said, okay, here's all my data. I want you to give me the pessimistic take on what I'm facing here. And the chatbot said, oh, you have lots of structural heart disease. These arrhythmias are a warning sign for danger. Your 10-year survival is much lower than an average man of your age with your health status.

John Lantos

I said, give me the optimistic one. He said, you know, you've been asymptomatic. Your ejection fraction is a little low, but it's not in the danger range. Your chances of living another 10 years are excellent. If you take good care of yourself, take your medicines, keep your blood pressure under control and your, you know, lipid profile looking good. So we now have the opportunity with these LLMs to design our own doctors and get not just one opinion, but as many different opinions as we want. It's an important process because it's one that requires sophistication in our interactions with the chatbots. But it's revolutionary, I think. It is going to lead to a new understanding of what patient autonomy means, a new understanding of what shared decision making means, and new ways of thinking about empowerment. So, yeah, I think AI is transforming and will continue to transform lots of the ways that we think about doctors, patients, disease, and treatment.

John Lantos

The question that remains, what's the role of the doctor going to be in all this? And I think doctors are going to need to learn to work with this new technology and to help patients understand both its potential and its limitations.

Emily Silverman

I love this idea of answer the question as an optimist, answer the question, as a pessimist. And it reminded me of, I don't know if you've heard of this physician writer, Jay Baruch. He's an emergency medicine physician at Brown, and he wrote a book called Tornado of Life, which is a series of essays and vignettes from his time in the emergency room. And there's one chapter where he focuses on his own health journey, and he writes about something cardiac. I think he needed a valve replacement or something like that. And he interviewed two cardiac surgeons. The first one, he went with his wife, really nice guy, sat down, listened to him, explained everything, patient, answered his questions, knew all the information. And then they walked out and the wife was like, oh, he seemed really nice. I really liked him. And he said, well, let's meet with the other one and then see. So then he went to the other one. And he was arrogant, and he was a jerk.

Emily Silverman

And he kind of blew him off, just had an attitude. And, you know, just that was that was his vibe. And Jay and his wife walked out of that meeting. And Jay said, that's my guy. And he picked the asshole. And I had him on the podcast to talk about it. And I think I even asked him, like, why did you pick the asshole? And I don't know, he said, for whatever reason, I felt like that's what I needed in that moment. Like, that's what made me feel safe, for whatever reason, in that moment. And so it just made me think about GPT. And it's like, you know, talk to me like you're an asshole surgeon. Talk to me, talk to me like you're an empathetic pediatrician. I don't know. I mean, do you think that could confuse patients? Because if the spectrum of optimistic and pessimistic is too broad, do you just feel more lost? Or do you feel like it's empowering to be able to choose like what kind of outlook or personality you want to hear the response coming from?

John Lantos

Yeah, it's a it's a great story. I mean, and Baruch is a great writer. And going through the process was a clever way. I don't know if he knew that the doctors were going to be so different in their approaches. But it illustrates how idiosyncratic patients' needs and desires are, and how random the responses they get if they don't have the knowledge or will to go seek different opinions the way. Baruch did when he was facing his own health decisions. The question of how to evaluate it, and whether it's going to confuse people or help them, empower them or disempower them. One of the things that makes AI unique in the trajectory that I described of, you know, this 200 year long process by which we've been glorifying the dehumanization of medicine is that AI is available to everybody on their phones. Everybody carries this technology in their pockets. So my sense from talking to a lot of people in hospitals today is everybody's using it and nobody's evaluating it. And the attempts to evaluate it are running into major conceptual problems because we don't know how to evaluate a technology that is constantly changing by design.

John Lantos

Every interaction that you have with any particular large language model will be different every time you have that interaction because that's the power of the technology. It's constantly bringing in new information. It's constantly adjusting its own internal weighting system. And people have talked about this as the black box problem of the fact that we don't understand, even the people who design it don't understand how it gets to the particular conclusions that it gets to. This was like a revelation when, you know, the first chess playing AI or the first go playing AI made moves that no human had ever made before. And, you know, before that people thought of AI the way we used to think about primitive medical AI, like the computers that read cardiograms, they were programmed how to read them. They followed the program and they did a very good job. It was a defined task. It was discrete. You could measure how's the AI doing against the best cardiologist. How's it doing against the average ER doc because it was a fixed thing. No more. It's changing every time.

John Lantos

So how will we know whether this is leading to better health outcomes? It's going to be a problem for study design people and researchers to evaluate a constantly changing technology that's available to everybody and being used in highly personal idiosyncratic ways by both patients and doctors and sometimes patients and doctors together. It's exuberant. It's wild. And to my mind, it's almost unassessable.

Emily Silverman

Hmm. In the analogy of the aura, I mean, the title of the piece is the loss or the diminishment of the physician aura. Is it the case that these GPT interactions are like Warhol Campbell soup can prints and they're everywhere. There's different colors, but there's, you know, thousands of them. They're reproducible. They're cheap. They're democratized. And then the meeting with the doctor in the office is like the one of a kind van Gogh starry night that has the aura of like, you can see his brush strokes. Does that analogy still hold or do you feel like the AI has an aura now and the physician, I mean, I guess the physician is in a lot of ways losing the aura.

John Lantos

Well, I love that question. One take home lesson from Walter Benjamin and his insights and what's happened since he wrote that great essay is that he imagined somehow that film and photography would replace painting. And in some ways they have. And Andy Warhol is a great transition figure here since he played with the idea of what is art, what is art for, and is a Campbell's soup can a form of art. And if it's not in itself, can you make it into one simply by putting it through some sort of artistic transformation? What's clear is that painting and visual art outside of photography has not disappeared. Instead, in response to the challenge of technological reproduction, it's transformed itself and it's transformed its social role. So the aura no longer attaches to the sensitively done portrait or even something like Van Gogh's Starry Starry Night. Instead, it attaches to different schools of visual art, Impressionism, Cubism, abstract expressionism that have questioned the social function of art and changed the way we think about what it means to be a valuable, important piece of art in ways that are not mechanically reproducible.

John Lantos

In a way, it's enhanced the aura around original works of art. I think in medicine, and I talk about this a little bit in the piece, I say a similar thing is going to happen and we're going to have to question what's the social value of medicine if it can be cheaply reproduced like a Warhol print and everybody can have an expert clinician in their pocket.

John Lantos

What then is the role of doctors or hospitals or medical journals or what challenges will medical educators face as they try to train the next generations of doctors who are going to practice in a world where this technology is ubiquitous? Many of the things that we've valued in medical education, cramming as much information as possible so the doctor could be knowledgeable about cirrhosis and pneumonia and COPD, and that's not going to be necessary anymore. So what do we train doctors to do? And if we think humans have some unique capacities that the chatbots don't yet have or can never have, how can we foster those? How can we teach medical students to be the kind of doctor who can work collaboratively with this powerful technology? To bring about the best patient satisfaction and the best outcomes.

Emily Silverman

With the painting, photography, metaphor, I'm just thinking about my own personal relationship to visual art. It's been kind of a strange zigzag. I actually majored in history of art and architecture at Brown, just like you. I went to Brown. And I don't know, after college, I got into this weird place with visual art where I decided that I just wasn't that into it anymore. Like, I would go to a museum and I would walk around and there were paintings on the walls. And I was just like, I don't know, I think this might actually just be really boring. And I started getting more interested in other artistic mediums. I started getting more interested in literature and film and music. And in film, there is music, you know, like you listen to the Star Wars theme song or, you know, it's just so evocative. A lot of the movie soundtracks, you know, the music, it's almost like today's Beethoven is like John Williams. And you're watching a story paired with the music.

Emily Silverman

And it's just, it's just so rich and so evocative. And I remember thinking to myself, like, there's just no way a painting, a flat surface could ever convey that much depth or emotion. And so I got sort of distanced from visual art for a while. And then in the last few years, I came back and I realized it was actually a particular piece of art that I saw. I came across this painting of a, like a 12 year old girl and she was sitting on the ground and her knees were up and she was holding her knees and she was looking up and it was painted in a very expressionist style. Her, uh, her face was a long oval. Her hair was kind of long and she was looking up. Her eyes were looking up in this particular way. And she just looked so, um, sweet and innocent, but also scared. And you look at her and you can sense that she's looking back at you and she just wants to know that she's okay. Like she just wants to know that she's okay. And it's, it was a, it was the way in her body was elongated. It's all, it's hard to even put words to exactly the way that she was rendered in this abstract way was so emotionally evocative in a way that a photograph never could have been.

Emily Silverman

And then I realized that there's just a way to paint things that photographs can't access, that slant, like painting it with a slant with certain style or a certain color pattern. And so now I have a lot more respect for visual art again. And I'm just thinking about that journey for me. And I'm trying to map that on to like, what does that mean for the physician? I mean, I think for me, I love the feeling of speaking to an expert. Um, even just sitting here, speaking with you, 20 years of pediatrics, 20 years of babies in the NICU, 20 years of bioethics, like, God forbid I ever have a baby in the NICU. Like, I would love to speak with you about that and just be like, in your experience, what do you think? What have you seen? How do you evaluate my situation? Um, I'm sure I could ask GPT the same thing, but there is something to that personal history and that personal depth of experience that I think is irreplaceable. But I don't know, I don't know what you think about that.

John Lantos

I, uh, your journey fascinates me because in some ways it recapitulates the trajectory that Benjamin was describing, like the importance of an individual painting or work of art is going to disappear. Instead, you're going to have Star Wars with the music and the movies. But then rediscovering something unique about one particular painter's interpretation of a moment in the life of a child that somehow illustrated or captured an essence that no photograph could, and that must have reflected a deep connection between artist and subject. To me, one of the interesting things about the way the chatbots are able to be empathic, humanistic, and understanding is that they're being trained on works like that. Some of them in visual arts, but more we're living in a golden age of medical humanities. The amount of writing about what it means to have an illness, what it means to face decisions.

John Lantos

I mean, everything from the astounding popularity of The Pitt on television, uh, or the amount of medical themes in other popular TV shows that aren't necessarily medical like Star Trek or movies like Pedro Almodóvar's All About My Mother with its theme of heart transplantation or his latest, The Room Next Door, about choices to end one's life by medically assisted suicide. These analyses are delving into the deepest meanings of embodied frailty and mortality, the things that have always been what the primary purpose of medicine is all about. I mean, doctors or healers or shamans always had an exalted role in societies long before they had effective interventions or long before they had the kind of interventions that today we deem effective. What they had instead was a deep understanding of fundamental human needs as people were facing illness, as they were facing their own decline, their own mortality. The healer, the doctor, the shaman, the human assigned to the task of interpretation would come into that and give some meaning to it in a way that people found healing and comforting, even if it wasn't curative.

John Lantos

Medicine today is much more complicated and has many more interventions, but most of our interventions are halfway technologies. Often what we're doing is substituting a medical sort of frailty for a natural sort of frailty. I now take seven medications a day for my cardiac problems and the medications cause problems and it's hard to understand whether the medications are doing me more good than harm. I'm in an evidence-free zone, but as medicine gets better and more personalized, everybody's going to be in an evidence-free zone. We're all going to have our unique genomics that we are going to be able to analyze. There are going to be five, 10, 15 different options for treatment for every sort of problem, and we're going to have to make some sense out of both what's happening to us and what the world offers us as a way of coping with what's happening to us.

John Lantos

And this exuberant creativity in writing, in film, in drama, Broadway plays, memoirs, novels, is exploring that territory and the large language models are being trained on it. So I think we will find that those tasks, the sort of pre-scientific tasks of the healer, are going to become more and more important in an almost ironic way. That is, as the science gets better, the role of a doctor in applying the science will get less important because the better the science, the more it can be mechanized, reproduced at scale, and delivered without the need for a sensitive one-on-one doctor-patient interaction. What's left, though, will be when we've run to the end of what scientific medicine understands. But as patients, we're still dealing with all the choices that we have to make. And as doctors, we have to help people navigate that unexplored terrain. So in some ways, I think it offers opportunities, sort of like the opportunity that painter you love took advantage of to find a new way to portray what was going on in the world of that child.

Emily Silverman

Well, we've covered a lot of ground. I don't want to take up too much more of your time. This has been a really fun conversation. Maybe to end, you can just share with us anything coming up for you. What's next? Any more writing or what's on tap? What's on your mind? What are you thinking about these days moving forward?

John Lantos

Yeah, I'm thinking about two of what I see as the most fundamental challenges for humanistic medicine, trust and empathy. So with trust, one of the confusions about artificial intelligence and the chatbot, LLMs, whatever you want to call them, is how can we know whether they're trustworthy? Trustworthy.

John Lantos

To me, and I'll say the same thing about empathy, that opens up a field of inquiry about how do we know that humans are trustworthy? What does it mean to be trustworthy? And what does it mean to put your trust in another entity at a moment of the most profound human need? So coming to understand what it means to trust and what it means to be trustworthy is something that philosophers have been working on a little bit. But I think AI is going to be almost like a new sort of laboratory tool that we can use to explore the fundamental nature of trusting relationships. To the extent that AI becomes more trustworthy, we'll be able to query it and figure out what exactly it's doing, how it's conveying its trustworthiness in a way that can help us establish our own trustworthiness. Same is true for empathy. I mean, people talk a lot about empathy as one of the most important elements of doctor-patient relationships.

John Lantos

Some people are a little more skeptical, but I think most people would prefer an empathic doctor, although thinking of the Jay Baruch story. Empathy, there's always a trade-off between empathy and what's perceived as competence. I would assume if we had Jay on the show, he would say something like, I just thought the asshole was, I thought his arrogance suggested more trustworthiness or something.

John Lantos

With empathy, the question is, what does it mean when I enter into a conversation with you and at some point say, I understand what you're feeling. I understand what you're going through. And you perceive that as valid. Often when somebody says that, we don't perceive it as valid. And in many cases, people's response is, you have no idea what I'm going through. How could you understand what I'm going through? And this again is something that a lot of patients have written about. Aleksandar Hemon had a great story. He's a novelist whose little daughter had a brain tumor, and he called his story about the experience of being the parent, the aquarium. He said, being in the hospital and that sort of situation, he felt like he was behind glass, that the rest of the healthy world was out there and he was in here and they could have no idea what he was going through. Lorrie Moore wrote a story called People Like That Are The Only People Here, about a similar experience where suddenly she bonded with all the other parents, with whom otherwise she would have nothing in common because they shared this experience of crisis as their loved one was sick or dying.

John Lantos

How people come to understand what another person's experience is like in a way that that person experiences as empathic is a fascinating question. There have been some, to my mind, pretty superficial studies showing that people perceive AI as more empathic than humans, superficial because they're all typed and they're about one particular question, and whether the human doctor's response was as empathic as the chatbots. But we know doctors, that's not what most doctors and patients are facing there. In the midst of an ongoing story, the doctor's seeing many patients over time, people have multiple problems. And nonverbal language is, of course, as important or more important in conveying empathy as the words that you would type on the page. But what is that thing, that empathy thing? And how can we use AI to help us understand it better by examining what it is that people perceive as empathy and then learning how to do that thing? Hmm. I think we have a lot to learn and AI is going to help us.

Emily Silverman

Hmm. I love that. Well, let's leave it there. Thank you so much for joining me today. I've loved this conversation and I have been speaking with Dr. John Lantos. John, thank you again.

John Lantos

Thank you. This was great.

Emily Silverman

This episode of The Nocturnists was produced by me and producer and head of story development, Molly Rose-Williams. Our executive producer is Ali Block and Ashley Pettit is our program director. Original theme music was composed by Yosef Munro with additional music from Blue Dot Sessions. The Nocturnists is made possible by listeners like you. If you enjoy what you hear and you want to support our work, consider subscribing to The Nocturnists Plus. You'll get access to The Nocturnists After Hours, our monthly bonus series featuring original conversations from our team, along with merch discounts and a full archive of episodes all in one place. Subscriptions start at just $10 a month. Learn more at thenocturnists.org slash plus or through the link in the description. Thank you for listening and being part of this community. Until next time, I'm your host, Emily Silverman.