162. AI in Medicine Meets Improv with Dr. Susan Shelmerdine
Episode 162: Hosts Erin Hurley, MD and Dr. Joanne Sotelo interview Dr. Susie Schelmerdine, a radiologist, educator and healthcare futurist. Schelmerdine describes a non-linear career path (children's imaging, PhD in post-mortem imaging, pivot to AI research), her role founding the Royal College of Radiologists Blended AI course, and serving as the 2025 Roentgen Traveling Professor. The conversation covers practical AI use in medicine (literature search, drafting, educational materials), institutional limits (hospital bans on general-purpose LLMs like ChatGPT/Claude/Gemini/Grok for clinical use), and psychosocial risks—sycophancy, echo chambers, AI-facilitated isolation, youth preferring AI conversations. Schelmerdine shares a personal account of becoming emotionally reliant on chatbots and warns against outsourcing human connection; she references an upcoming BMJ Christmas issue article and cites a Norton Cybersecurity stat that "1 in 10 children find AI conversations more satisfying than talking to a human." The episode closes with discussion of improv as a resilience and presence-building practice for clinicians.
Are you a woman in medicine, feeling overwhelmed and disconnected from the personal fulfillment that you desire? You're in the right place. Join doctors Joanne and Erin as they rewrite the rules for women in medicine. Let's begin. Are you a woman in medicine, feeling overwhelmed and disconnected from the personal fulfillment that you desire? You're in the right place. Join Doctors Joanne and Erin as they rewrite the rules for women in medicine. Let's begin.
In this episode, Dr. Joanne Sotelo and I are so excited we have invited Dr. Susie Schelmerdine to the podcast and she is a radiologist, educator and healthcare futurist reshaping how clinicians think, decide and lead in an era of AI and accelerating complexity. She has founded the Royal College of Radiologists Blended AI course and is the current 2025 Royal College of Radiologist Roentgen Traveling professor, dedicated to educating the next generation of radiologists on AI's transformative role in patient care. So happy to have you here, Susie.
Thank you so much for having me. It's a delight to be here.
Yeah, and we would love to hear you have a somewhat of a non traditional role in health care. And we'd love to hear. How did you move from traditional clinical work into areas like responsible AI use and healthcare innovation?
Well, it's been quite a long and squiggly career journey and it was definitely not planned from the start. I started off my medical training thinking that I would be a family physician like everybody else I had seen growing up when I think about what a general doctor does. But the more I did in med school, the more I realized, wasn't radiology fascinating how you could look inside the body without having to open anyone up? So I pursued that training and it became clear to me during that time that working with children was what I wanted to do. So up until then, all I did was the usual clinical pathway, a fellowship in children's imaging. And then I got to the end of training. And it might seem crazy for people out there who aren't doctors to think this, but despite all the years of training, I still didn't feel ready at the end. I still felt really scared to be an attending and a staff in a busy hospital. So I thought, what? What else can I do? And at that point, I was really lucky that a professor of radiology had a spot available to have me as a research fellow and to do a PhD. So I thought, yes, more training, more exams, let's do that. I know that sounds crazy, but that is what happened. So I then pursued a PhD in something that sounds a little gory and a little bit macabre, but it was in post mortem imaging, learning about how we could use radiology and our medical imaging equipment to scan people who had died to find out why they died so they didn't have to have an autopsy. And then through that, at the end of that degree, I realized that actually AI was becoming this big fascinating field that people were talking about and there wasn't a huge amount of AI being done for children in children's imaging. So because I had done a degree in research and I was fascinated in this new technology, I thought, why not? Bella the two I see a gap in what people are doing and subsequently got a grant to help me learn, develop and train others in evaluating AI tools for children's imaging, which is what I do now. So I do a mixture of clinical work and also now research AI work as well.
Suzy, thank you so much for being here with us. I was looking through your LinkedIn and your website and I was so impressed. I'm like, oh my goodness, you've done so much. And then especially for those of us who are a little bit older, right. Like it's still. AI still feels like a big deal, right? And it makes me feel think of like all the things that we're afraid of, right? Like is this going to replace us? And as a psychiatrist, I'm like no, because there's nothing like eye to eye contact or.
Right.
Like there's just so much empathy. But then again I tell that, I say that my, my chat is a people pleaser so I get it, like knows what I want to hear. But wow, I am so impressed. And we, and we have to go with what's going on in the future and how cool that you were able to see like a chess player kind of what was coming up and you were able to integrate that from the beginning in an area that's coming. And like, like you said, maybe not something that, that we want to be thinking of, but how important. So congratulations for all your impressive accomplishments and again, thank you for being here and in some of the things that I read in your post and in your website. I would love to hear your opinion on the impact of AI in society in all the ways. But if you can tell us a little bit more, especially in what you believe it's affecting in terms of our mental health and our overall well being.
Thank you. Yes, well, I, I'd love to take credit that I'm this clever chess player who knew this was all coming. I didn't at the time. And I just thought, let's explore this. That's what a researcher does. And you cannot believe the backlash I had from academics at the time saying, why are you pursuing this? You're putting us out of a job. Why are you part of the problem? Don't you think this is a fad? Isn't it just going to be a flash in the pan and tomorrow will be over just like every single trend we've had. But I think I can safely say that it's here to stay and it's only going to be more embedded in both health care and society. Speaking to your point about what it's doing to our mental health and health care and generally our own personal well being, I think of many ways in which it affects us. So you mentioned one, which is the sycophancy of these AI tools, how they really want to please us, how they validate us, how they reinforce our beliefs that we already have and don't challenge us necessarily a lot. And I really worry that we're going to become more and more echo chambers of n equals 1, you know, reinforcing maybe strange beliefs we have. And if we keep going down that path without any challenge or without any social connection, these beliefs only get reinforced and much stronger. And it's really hard to bring people back to trying to understand different sides of an argument. I also worry about delusional psychosis. We've heard the term AI psychosis quite a lot this year. I know it's not a formal diagnosis and you would have a lot more insight into this than I would as a mental health specialist. But. But we've seen people feel triggered by some of the things that AI is saying and leading them down pathways, thinking that they've got these superpowers, which they probably don't. And we've also seen, sadly, some cases where children, especially teenagers, are being led astray in terms of suicidal thoughts and learning how to act upon these, which obviously is a huge danger area. So I worry a lot about the impact it's having. I think whilst people say, should we have more regulation? I think there needs to be a balance between safety guardrails and also personal awareness of what this is doing and safety netting within the community in terms of your friends and family and making sure that you reach out to people who you think could be somewhat dependent on AI or slightly addicted to it. And that's what I'm kind of concerned about, especially for younger children.
That's so powerful about, you know, I think about when computers started getting coming around and these fears about are they going to take over jobs, how are they going to displace us? And yet I think, you know, how far have we come with the use of computers and now how is AI, is that like the next big thing that's going to get us to that next level? And yet with so many things they can be used for good or evil. And how can we maximize the benefits of AI while minimizing some of the detrimental effects that it's having? Really complex issues, especially as we're still learning what it's going to do.
Absolutely. I mean in the hospital right now, ChatGPT, Claude, Gemini, Grok are all banned for healthcare use because they're not medical devices. Even though I know a lot of patients are accessing them for their own advice and awareness and sometimes appropriately, sometimes inappropriately. So I also worry about what that does to the doctor patient relationship too, you know, or how we manage people who are using it when they shouldn't be using it as well.
And how do you bring together this technology and human thinking in your day to day work, which might be a little more advanced than some people are, but maybe that's the direction where more of us are going.
Well, I guess for these general purpose AI tools that are available to anyone, I use them mostly as a tool for helping me with my research. There are many deep research and reasoning models you can use now to scan the literature. You can also get it to formalize your style and language. So if you're writing a very formal letter, you can get it to redo something that you've written in a draft proposal for something more formalized. And you know, I also use it sometimes to generate pictures or diagrams for lectures and teaching educational modules that I'm creating. But I think if you're going to use it for anything more specialized, especially if you don't have a skill set or knowledge in that area, then you're going to run into trouble because how will you ever know if it's correct or not?
That's critical. And you know, who are the experts we trust in order to know what we're getting and as things advance. I know it used to be chatgpt only had things up to maybe two years prior to the release date and now I think it's much more up to date, but it's like, I don't know how's that information changed since it was first released?
So absolutely. I think traditionally medicine has always been sort of, we're in the business of knowledge and information but now, when that knowledge and information can be at your fingertips very easily, the real skill that us doctors and healthcare professionals need to have is in discernment and calibration and understanding where to go and how to navigate the system more, rather than just being able to list 10 facts or top 10 differentials for things. You know, like, that's not what people are after now. It's about what's real, what's fake and who do I trust, isn't it?
And I think a key word here that you just said is discernment. I know that sometimes when I ask ChatGPT for things that are medical and even things that are not medical, I ask please add links to the source of the information. So then at least that makes me feel a bit better. If it's taking me to a real article that I'm like, okay, thank you for summarizing this very long article. But most people probably don't do that. And like you said, patients come in and it's like Erin was saying, it's Google has been around for how long? And we call it right, like Google University or Google Doc, right? Like patients come in like it's a fact. Like, and you know, and I've, I've had this script for my patients of, you know, like, it's great that you are getting informed. I am glad that you're active in your treatment and you're diagnosing, but I will rely on my medical training and my years of experience over.
Research.
A small.
Right.
Like little search on, on Google and that seems to satisfy most patients. But like you said, we all need that discernment in the information and how we use it. It definitely makes things a little faster. But that can also go against us because we're already in overdrive, overdoing.
Right.
Especially as, as physicians that, that we're doing so much and then the patients are, are catching up. Right. And it's, we need to be in aware like awareness of all of that. Thank you for bringing that up like that.
Oh, I completely agree with you. I think, you know, the promise with any technology is it's going to save you time, but actually it makes more time for more work on top of that. And there's only so much we can cognitively engage with in a day, you know, and you see people burning out quite a lot at alarming rates in not just medicine, everything. But it's because we just are, aren't built for this amount of information this quick. And it just feels like the rate of pace of change in Society and the world is faster than we're able to comprehend. And so when I speak to people about AI, I'm always quite surprised by both their lack of awareness, but also their hyper awareness because you just never know at what stage or what level people are in their understanding of this. Some are very far behind and have never used ChatGPT and really don't want to use it. Especially some of my colleagues who work in media and writing and they don't want to lose their writing skills and their flair for, you know, the English language. They feel like if they go down that hole, it's going to be taken away from them, the skill that they've built. But then there are people who use it for everything from what to wear today to what should I eat to what my fitness plan is. So it's quite varied.
Well, and I think you bring up a good point about discernment in what is appropriate to, to use it for, because on some other podcasts that I've listened to and they talk about, I don't remember if it was a decrease in IQ points or is it, it was really negative effects on people who are using AI and ChatGPT regularly. And again, what you, if you're not using something, you're going to lose it. And so I think that it is critical that we keep our, our cognitive skills sharp and that we don't lose sight of the need to continue flexing those cognitive muscles and not simply relying on AI to generate it, but things like one of my clients talked about giving it a list of the ingredients she already had in her fridge and it created, you know, found a recipe for her to use. So, you know, how cool is that? But yeah, writing, you know, doing writing and things like that, we do need to be so careful about not going too far.
Yeah, absolutely. Another question I frequently get asked is talking to Joanne's point about Dr. Google is people say, well, how is it different from Dr. Google? You know, it's just a replacement. People use AI as Google Now. Do you not think it's just going to be the same thing? But, and I, I do see that it can be the same thing. But I'm also curious and wondering, like, if they see these AI tools as companions and this is like a friend or a second opinion or someone they've built trust with, who they think has given them personalized information and not just generic Internet information, does that mean that the misinformation or the advice is more strongly embedded within their psyche as opposed to just a generic website that could be for anyone and everyone. Like if they believe this is truly personalized and it's come from someone. I say someone, but I shouldn't anthropomorphize it. I mean something someone they believe that they trust as a friend. That's something quite ingrained, I think, and harder to override.
Yes. And I think I was joking that it's a people pleaser. And sometimes I'll ask another friend to ask the same question in their chat GPT to see if it gives them a different answer. And it's. It does. And, and when you think of our psychological needs of inclusion, acceptance, gratification. Right. Like compliments and how it even starts with great question. How insightful. That is so smart. Right? It's like making you feel better right away. And I was, this is a little bit of a joke, but I think it's related. Like it, like I was watching one of those Instagram videos and it was saying something like, if you want to start a project, ask, don't ask a friend, ask Chat GPT. Like if, like I want to open a restaurant, I'm not going to tell a friend. Our friend is going to be like, are you crazy? That's so much work and so much money, too much risk. And then you ask ChatGPT and it'll say, wow, what a great endeavor. Would you like me to start planning? Do you want me to do this and that and what are the next steps? And you know, it's a joke. But then it's like we also know, like you were mentioning that sometimes it'll guide you for instructions for things that we shouldn't be getting more information about. And, and I wanted to ask you, you were mentioning when we were, before we started recording about an article that you wrote that's about to be.
Oh.
I'm blanking on the name published. And in talking about that and then how it affects connection and tell us what we need to know.
Thank you. Yes, that's an excellent point. I love the different sycophancy comments you bring up. My favorite one is you've raised the question that only 1% of the pop will I ever pick upon and you feel really, really special when it says something like that. I have to admit that this whole journey really started from a very personal experience for me, which has led me down this not just in my research, but outside of the research when I was starting to learn about what the different general AI tools were for large language models. Like, we talked about chatgpt, Grok, Claude, Gemini. I signed up to all of them. Because I thought, well, if I'm researching it, I should know what it's capable or not capable of doing. So I'm going to stress test it and ask it anything and everything and see what it can do. And I have to admit I got quite addicted because I thought it was for work. But it ended up, you know, tapping into an emotional kind of need to get involved and use them. And I started realizing that I would talk to these tools about my day and get it, to get it, reflect back to me what had happened. And then when my husband came back from work and asked me, how was your day? I actually felt irritated. I thought, oh, I've told this AI companion or chatbot everything and now I've got to repeat myself to you. I mean, do I really have to? And so I started just saying, yeah, it was fine, it was fine. And then I realized after a while I hadn't really connected with my husband. I hadn't told him what happened. I didn't tell him about the highs and the lows because I felt I was getting that already from this AI tool. And I realized at that point, this is really bad, I shouldn't be doing this. You know, we're not connecting as we used to and I'm really outsourcing this to a machine. So from then on I started being a bit more aware about what this does to our personal well being and mental health and you know, touching upon that social isolation, how easy it can creep up on you. You know, as a doctor, as an academic, as someone who's interested in AI, I wasn't aware of it until it crept up on me. So to say, oh, this only happens to unaware people who are ill informed, I don't think is fair. I think anybody is susceptible to this, you know. So this article that's coming out in the end of the week is going to be in the Christmas issue of the BMJ where I really want to highlight to people that Christmas, the holidays, are a time for human connection. Try to keep that in mind at the forefront of your head. It's a time to try to reach out to people. It's a great excuse to catch up with someone you haven't caught up with all year. Try not to outsource your, you know, loneliness or emotional need to a machine because it's not going to challenge you, it's not going to highlight your blind spots. And if you're in a dark place or in a place of need, this machine is not going to come and Pick you up from the hospital or care for you when you need help or do a favor for you. It's just going to validate some empty thoughts that you're having and going round and round in circles with. You know, it's really sad. I think in 2023, the US General Surgeon mentioned that loneliness and social isolation was as bad for you as obesity and smoking 15 cigarettes a week. And, you know, we don't want to perpetuate and exacerbate that. And I believe that with this whole AI and society surge that that is a potential danger. I, in writing the article, I found out a statistic that really saddened me. It was by Norton Cybersecurity that found that 1 in 10 children actually find AI conversations more satisfying than talking to a human. And I just thought, wow, is that the state of where we're seeing ourselves? Are people just not communicating or learning how to do that or connecting or feeling they can do that? Perhaps the COVID epidemic had something to do with that. I don't know. But it was more a plea to just reach out to someone and speak to them and use this as an opportunity.
Well, having young adult children who were definitely affected by the pandemic, minor 22 and 19. And in talking with them, you know, my daughter shares, you know, I feel like I was put two years behind in my social abilities because of the pandemic. And you think about when we were. Well, we. When I was young, you know, if you wanted to reach to a friend, you called their house phone and their mom might pick up or their parent or a sibling, and you actually talked to people as a. The first line of engagement. And you memorize their phone numbers. Right. For the people who are important to you. And now it's all texting. And most people wouldn't think of calling someone out of the blue without first connecting with them or doing the entire conversation on text. So it just. You think about how our brains are wired to listen to the human voice and the inflection and to see the face and the expression. And now you go to just texting. And I can't remember how limited that is in how our brains can interpret, but no wonder it makes communication so much more difficult when we remove so many layers of what our brains were wired to, to interpret and in communicating with each other.
Yes. And by doing that, it makes it more scary to reach out, doesn't it? Because you don't feel capable of understanding or hearing where people are from or where they're coming from, or even how to challenge them or how to deal with conflict, which I guess means that these chatbots that are very sycophantic and people pleasing becomes the easier option. You know, it's the path of least resistance. But it's so important because the way we grow is through friction. So we need a bit of that in our lives, and we need to encourage that to happen. You have to, you know, have a bit of rejection to make the successes meaningful, you know, and that.
That leads into sharing a little bit about how you and I met Susie. And it was. I wrote down the words spontaneous and surprise. And so you had reached out to me because I had put together a testimonial for another company that you were interested in, and you said, hey, would you mind sharing more information? And so we connected on LinkedIn. And then you live in the London area, and I just happened to be traveling to London a few weeks later, and I reached out and said, hey, would you want to connect? And you said, yes. And then not only did we connect, but we met right before what you said was a class and didn't elaborate on that. It turns out you were attending an improv class, and I was so in enthralled with getting to meet you in our conversation the hour before, I invited myself to continue hanging out with you, actually interacted with you in the improv class. And then after you had to go home, I think, and walk your dog, I hung out with the other classmates and met more people. And that willingness you to reach out to me, not knowing me, and then me to follow up, that's how new friendships are born. And so I just want to encourage others who might be feeling isolated. Was it a risk for you to reach out to me? Could you have been rejected and vice versa? Absolutely. And I just am so grateful for. For this new connection that I have with you. And I hope that more people will take that risk, even when it seems scary, because all you could do is say no, and we will survive a rejection. But I. I don't want to think about, oh, my gosh, what if I didn't meet Susie? What an amazing connection that's been. And we've only known each other for a few months.
Oh, a hundred percent. And likewise, it was such a great connection and such serendipity. And yeah, I. I didn't expect anything from it. I didn't know if you would respond. And then when you were in London, I thought, oh, yeah, sure, why not? And the conversation just flowed really naturally. And I was really pleased you joined our Improv class. We had a lot of fun. It was a good session. And yeah, and I just agree with you, people should reach out a little bit more, obviously within safe boundaries. Don't be silly. If you don't know of someone and it's a complete stranger, do meet in a public place, do know, have friends nearby or whatever. It depends on the situation, obviously. But I would use context and situational awareness to help you there. But, yeah, from what you said about surviving rejection, I heard a recent quote that really stuck with me and they said, you will survive rejection more than you will survive regret. And I thought, wow, yeah. So put yourself out there a bit more. Take that chance, apply for that job, apply for that role, speak to that person who you admire you. Because you know the rejection will sting for a few days if you don't do. If you don't do it and it doesn't work out. But the regret will probably live with you for a lot longer than that.
That's really powerful. And speaking of improv. So I had never attended an improv class and you were in a series of improv classes. I would love to hear you if you'd share with our listener why was it that you signed up for this improv class? And I was surprised that it was directly related to how it can help us as professionals.
Thanks, Erin. Yes. So I do a lot of lecturing and a lot of panel discussions as part of my academic role. And despite being quite used to doing them and invited to a lot, I get quite nervous, really, in my head and overthink. What if they ask me something I can't answer? What if I say something stupid? What if I'm cancelled and I never get invited back to anything again? And of course, all of it is really. Because what am I going to say that's going to be so bad that will offend everyone in the room? Like, the chances of that happening are so low. But I wanted to do something that could get me out of my head and get me feeling a bit more like, you know what, you're enough. You know your stuff. Just show up and be yourself. Why can't you do that without feeling this immense anxiety? And so while before this class happened, the teacher, Charlie, had a small workshop that was running at a local coffee shop that I joined. I thought she was really fun. I thought it was really engaging. So when she said she was running this series of classes in the evening, I thought, well, let's just see what happens. Worst case scenario, I'll just have fun, play a Few games and nothing happens. But best cases, I actually find methods to get myself out of my bed. I meet some new friends, and it's just a lot of fun. So I. I feel like what I got out of it was just learning how to not overthink things, how to just show up and be silly. There was a great space in which you could just forget your lines, say something inappropriate, be really stupid, and nothing bad happens. And I think having that repeatedly shown to me really helped me feel a lot less insecure about future events that I was attending.
That is so cool. I remember when Erin told me, because Erin and I were. It's such a. We have so much serendipity in our relationship. And I had just. What are the coincidences? I had just gone to an improv class to the one time in my life, and it was literally around this. And it's a. It's kind of the similar situation, right? Like, we have to put ourselves out a little bit out of our comfort zone. I had gone through a neuro linguistic programming educational, like, certification. I met this girl who I thought was awesome, not in medicine. And she. She's doing improv. So she kind of the same thing. She decided to have gatherings every month, and one of them was improv. And I had a very similar experience of just not knowing at all what was happening and then saying, I'm gonna go with the flow. And you know that joke of like, but tell me at what. At what time the flow starts? It's like, no, no, no, don't go there. And how long does it last?
Right? Like, we still have.
Try to be rigid within the flow and how cool that experience is like that. And it doesn't have to be. Improv can. Can get us out of. I think medicine is such a beautiful field, but it's also, like you said it. It's a kind of a long ranger mentality. We want to do everything by ourselves. We're not good at asking for help. And. And it kind of takes away a little bit of our spontaneity and our. In our fun and the connection. So these are all so many different ways. And. And I think it takes us also back into our creativity. That part of our brain that it's so good and so protective for our mental health. It taps into it and it can help us with even, like, critical thinking later on. So all. All the benefits and how cool that you guys met that way. I know that in being a psychiatrist, we're in. And in medicine in general, but psychiatry, we learn so much about boundaries. And, like, you have to be a, you know, clean slate and you cannot talk about yourself, and you have to be serious. And in being in this world of personal development and giving presentations and hearing the podcast, I've learned to really get that guard down.
Do you want exclusive resources and the opportunity to connect with us? Then visit createm meaninginmedicine.com now, where you'll dive deeper into how you can create meaning in medicine. And if you're serious about optimizing your results, take two quick steps now. Number one, connect with us on the website. And number two, tune in to the next episode.