Digital

ChatGPT in medical education: Generative AI and the future of artificial intelligence in health care

. 12 MIN READ

AMA Update covers a range of health care topics affecting the lives of physicians, residents, medical students and patients. From private practice and health system leaders to scientists and public health officials, hear from the experts in medicine on COVID-19, medical education, advocacy issues, burnout, vaccines and more.

Why should AI be in classrooms? How do you integrate AI into a curriculum? AI tools for medical education, using AI in medical training and how AI can help you study.

Our guest is Marc Triola, MD, associate dean of education informatics and director of the Institute for Innovations in Medical Education at NYU Grossman School of Medicine. AMA Chief Experience Officer Todd Unger hosts.

Speaker

  • Marc Triola, MD, associate dean, education informatics; director, Institute for Innovations in Medical Education, NYU Grossman School of Medicine 

AMA Recovery Plan for America’s Physicians

After fighting for physicians during the pandemic, the AMA is taking on the next extraordinary challenge: Renewing the nation’s commitment to physicians.

Unger: Hello and welcome to the AMA Update video and podcast. Today, we're talking about the ways in which AI is changing medical education. I'm joined today by Dr. Marc Triola, associate dean of education informatics and director of the Institute for Innovations in Medical Education at NYU Grossman School of Medicine in New York. I'm Todd Unger, AMA's chief experience officer in Chicago. Dr. Triola, welcome.

Dr. Triola: Thank you. It's great to be here and thanks for the opportunity.

Unger: Well, it seems like a lot longer than a year, but it has been about a year since ChatGPT got everybody talking about AI. At NYU Grossman, you're already using this technology to support students in brand new ways and we're eager to hear more about that. But, first, can you tell us what inspired you to leverage AI at your institution in the first place?

Dr. Triola: Well, I think like everyone around the world, we were struck by the power of AI, the rapidity by which these innovations were being developed, how accurate and realistic the text they were generating was, how well it handled understanding concepts in text, and we saw an application for AI to improve the work of our students, our faculty, our clinicians. And as an as a health system, we're really embracing AI across all three of our missions—our mission to teach medical students and residents, to care for our patients, and to discover new science. And we think that there are a huge number of applications to treat many of the pain points in, not just medical education, but the delivery of care.

Unger: It's interesting because so much of the discussion over the past year has been about concern about how students are going to use AI, so we're really talking about a different thing. So let's take a closer look at the ways that you're using AI, starting with how it's helping your students learn and access research. Tell us more about that.

Dr. Triola: Sure. One of the first projects that we did was a project called DX Mentor where we are using AI to automatically identify learning resources and the latest medical literature based on the diagnoses of the patients they're caring for. We built a system here that connects data from our Epic electronic health record, and it understands which patients in Epic are being cared for by our medical students, connecting that data with a digital catalog of educational resources—some of these things are infographics, are videos about clinical diagnoses and the medical literature itself—all of the latest guideline and review papers that are published in PubMed.

This system every day curates a custom set of nudges for each of our students that delivers to them recommended learning and recommended readings for the patient they had admitted the day before or the night before. It's just in time, it's curated completely by the AI—no human intervention needed—and it's customized for every student. So it's a nice example of precision medical education enabled by the power of generative AI and artificial intelligence systems.

Dr. Triola: Wow, really interesting and so practical to their experience with patient care. Another way that you're using AI is to help students study. I'm sure that's welcome news to your students. How does that exactly work?

Unger: Sure. And—so we're experimenting in this area and I want to point out that medical students across the country are experimenting with this. And medical students have been doing a fantastic job of organizing themselves, sharing prompts for the AI to generate USMLE-type questions, flashcards to study. They're even using the AI to create virtual patients that they can interview and diagnose and get feedback from the AI on their performance in gathering a history or being empathic with patients.

We've created a similar system here that curates publicly available data sets of self-study questions, the educational resources that I mentioned before, to connect the dots between what our students need at the individual level and this growing amount of digital educational resources. And the glue that can tie them all together so efficiently these days is AI and that's—so we're working very hard to implement that to really provide our students with a very precise and personalized suite of assessment and curricular items to help them learn and study in as frictionless a way as possible.

Half the dues, all the AMA benefits!

  • Free access to JAMA Network™ and CME
  • Save hundreds on insurance
  • Fight for physicians and patient rights

Unger: Gosh, I love that idea of the virtual patient and diagnosis. What a great example of that. In addition to all the studying, of course, med school is also a times for students to think about their career goals. And I know, having talked to a lot of students, what's on their mind the moment they get into medical school—great news is that they can now use AI to help them with this task. What can you tell us about how it's being used there?

Dr. Triola: So the use of AI to support our students with their goal setting and achieving their goals was an area that was surprising and powerful. Our students author goals that they share with their coaches in a custom software application that we've built here to support our students to help our faculty coaches over their time during medical school. And we've added ChatGPT to that as well.

When our students are writing goals, they have the ability to click a button and get feedback from ChatGPT suggestions on how they might accomplish that goal. And what we found is that by no means do the suggestions from ChatGPT replace the advice that they get from that faculty member who knows them well, but they are often very concrete and detailed suggestions—which journals should they read if they're interested in a particular a specialty.

ChatGPT even knows all about the clinics at our hospital and can suggest individual faculty at NYU or clinics that a student might want to shadow in or do research with if they want to learn more about a particular topic. So in this case it's been a nice supplement to our human coaches and a way that AI, at least in the education sphere, can begin to really enhance the ability of our faculty and our students to better leverage their time and make decisions quickly, and efficiently, and accurately about their career planning.

Unger: Really interesting. I think the word that is kept popping up as we've talked about AI with several folks over the past week has been about augmenting. Is that how you see it playing a role at your institution?

Dr. Triola: Absolutely, yeah. AI is about giving everybody superpowers. It's not about replacing doctors, it's not about replacing medical students, it's not even about replacing the skills that they need to learn. It's about empowering everyone to be at the top of their game and to improve the quality, efficiency, consistency of the work that they are going to be doing as people, as humans. And so really that that's our approach. This is about giving everybody superpowers.

Unger: I love that. And it's so refreshing to hear practical applications, the way that you're helping students in their experience because so much of the discussion around ChatGPT or AI is, at least on one end, about uncertainty and on the other end fear and so, wow, what a lot to have accomplished in such a short period of time. I'm curious what—as you look ahead, what's the next big opportunity to incorporate AI into medical education?

Dr. Triola: Well, I think AI is going to really help us accomplish this mission of precision medical education. We want to deliver the right education to the right learner at the right time. And this is a mission that we are happy to partner with the AMA in. This has always been difficult because having that customization for every student takes a lot of work and the AI systems can really take much of that work off of our plate.

AI in Health care CME on AMA Ed Hub™

Artificial and augmented intelligence are rapidly changing technologies with wide-reaching medical implications. Earn CME credits while learning at your own pace.

So I think in the future we're going to see medical education models that have AI as a co-pilot sitting next to the student, sitting next to their faculty and coaches, providing guidance and advice along the way, curating curricula and assessments, and really tailoring to—at the level of that individual student tailoring what they learn and how they learn so that they make the absolute best use of their time. And it's not about skipping tasks or a shortcut to clinical competence, it's about efficiently tailoring medical education. And it's about meeting the students where they want to be in terms of their goals. And this is going to be a really key way to do it.

I'd say the second point that is critical, that's going to change is that we're going to teach medical students how to—need to teach medical students how to use AI clinically. This will become a key part of the practice of medicine, whether it is AI to diagnose patients, AI to communicate with patients, or AI to actually write some of the notes and do clinical documentation or other things that we haven't even imagined yet. So creating students who are prepared for that future, who understand the powers the limitations, when they should really question or really think very carefully about the recommendations coming from AI is going to be really key. And this is another area that schools are really coming together, thinking about what that curriculum should be, and how we should integrate AI in the curriculum from that clinical perspective.

Unger: Now, it's interesting, you're in year one of this, you're already making these kinds of advances, you're talking about how they're the education of today's medical students is going to differ. Project yourself out 10 years and now you've had successive generations of incorporation of this technology into medical education, do you think it's going to look really different 10 years now—from now than it does currently?

Dr. Triola: I think it will look evolved—significantly evolved from where we are now. There have been many trends in medical education over the past decade that have been slowed by the pace by which faculty, by which the schools themselves have been able to implement these things. They've been slowed by our ability to perform medical education research to understand what approaches or techniques work best at producing competence and excellence in our students. And AI is going to speed all of these things up.

So I think that when we think about accelerating change in medical education, which is an initiative from the AMA, and we think about the initiatives from all of the different medical schools that are going on towards competency-based medical education, towards different models of medical education that aren't necessarily of the same fixed length or duration, that are really focused on graduating students who are at that level of competence that they were aiming for. I think that AI is really going to get us there. Although, as you point out, it's only been a year. So maybe what we'll see 10 years from now is something that is quite dramatically different.

I think one of the biggest important points that we should recognize is that 10 years from now the students entering medical school will have had the benefit throughout all of elementary school, high school and college of AI. And so that may dramatically change how and what our students are learning before they're even getting to us, which could simply change the whole baseline of where medical education is. I'm hugely optimistic and excited about this. I think this is really going to unlock many of the changes and many of the opportunities that we've been trying to strive for for years.

Unger: That is a really exciting vision and so interesting to hear your perspective and the progress that you've made already at your institution. Thank you so much for joining us, Dr. Triola. And we'll look forward to hearing more about your next set of innovations.

If you're interested in accelerating medical education and ensuring that technology is an asset and not a burden to future physicians, then join the AMA and make sure that happens. You can do that by going to ama-assn.org/join. That wraps up today's episode and we'll be back soon with another AMA Update. Be sure to subscribe for new episodes and find all our videos and podcasts at ama-assn.org/podcasts. Thanks for joining us today. Please take care.


Disclaimer: The viewpoints expressed in this video are those of the participants and/or do not necessarily reflect the views and policies of the AMA.

Subscribe to AMA Update

Get videos with expert opinions from the AMA on the most important health care topics affecting physicians, residents, medical students and patients—delivered to your inbox.

AMA Update podcast logo

FEATURED STORIES