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.
Featured topic and speakers
Is AI good for health care? How are doctors using artificial intelligence? What is AI used for in medicine? What percentage of doctors are using AI? Do doctors trust AI?
Our guest Margaret Lozovatsky, MD, vice president of Digital Health Innovations at the American Medical Association, discusses key findings from the latest survey on physician attitudes on AI in healthcare. Highlights include AI tools for administrative tasks, the importance of feedback loops, and AMA programs supporting physician AI education, artificial intelligence governance, and AI collaborative initiatives. AAMA Chief Experience Officer Todd Unger hosts.
- Watch AMA Update for health care news for physicians, residents and medical students.
- Not a physician? Join the Patients Action Network.
- Read AMA Augmented Intelligence Research on physician sentiments around the use of AI in heath care (PDF)
- AMA's Governance for Augmented Intelligence toolkit.
- AMA STEPS Forward® login.
- The AMA is your powerful ally, focused on addressing the issues important to you, so you can focus on what matters most—patients. We will meet this challenge together. Join us.
- Learn more about our AMA advocacy priorities, including:
- Reforming Medicare payment
- Fighting scope creep
- Fixing prior authorization
- Reducing physician burnout
- Making technology work for physicians
Speaker
- Margaret Lozovatsky, MD, vice president, Digital Health Innovations, American Medical Association
Transcript
Dr. Lozovatsky: This is when I say that the tech is there. It has moved way faster than we, as humans, have moved. And so over the next year or two, it's going to be time for us to slow down and really thoughtfully integrate.
Unger: Hello and welcome to the AMA Update video and podcast. Today, we're talking about the results of a new AMA survey on how physicians feel about AI and whether AI cares. Joining us today to discuss the survey is Dr. Margaret Lozovatsky, vice president of digital health innovations at the AMA, who's calling us today from Charlotte, North Carolina. I'm Todd Unger, AMA's chief experience officer in Chicago. Dr. Lozovatsky, welcome back.
Dr. Lozovatsky: Thank you for having me.
Unger: Well, this is the second year in a row that we've conducted this survey. I think a lot has changed, and we're going to talk about specifics. Number one, it seems physicians are feeling better about AI this time around. Let's hear more about the details.
Dr. Lozovatsky: You are absolutely right, Todd. That is exactly what we're finding. As you mentioned, we did the survey in 2023, and then we repeated it again in 2024. And what we're finding is that there's a growing confidence across all of our physicians in these tools. What we found is that 35% of physicians this time around said that they're more excited than concerned about the use of AI. And more importantly, what we saw is that many physicians felt that there was a definite advantage in patient care.
And those numbers were higher, again, than they were in 2023, which shows to me that physicians are starting to utilize these tools. They're starting to see a positive impact. And that impact is to our patients, which is really important and exciting as we look at how AI is being integrated into health care.
Unger: I've got to imagine that a point that you just brought up, which is they're using the tools more actively, as what might be at the heart of feeling better about it. What did the survey have to say about that connection?
Dr. Lozovatsky: Yeah, that's a great point and exactly what we found. And that was the result that showed the biggest growth. And so, in 2023, we saw that 38% of physicians were using AI tools, whereas, in 2024, it was up to 66%. And as I was thinking about that—I've been doing health care informatics for a long time—I have never seen an increase in usability that moved so quickly within one year, other than, perhaps, COVID. And then, of course, that usability went down, whereas this has been sustained over the last year. And so that was really interesting.
We also dug a little bit deeper into these numbers to understand what are some of these AI tools that the clinicians are using. And we found that most of the utilization was in the documentation space, some in note creation—in fact, most in note creation—but also in discharge instructions and care plans and tools for patients, as well as some translation services utilization, which is an area we don't talk about as much but is incredibly useful to help promote health equity.
Unger: Now, I've had a number of different conversations about AI with different physicians over the last year. And one thing that seems to keep coming through is that the tools are being used to reduce burdens. You mentioned note-taking. Is that a different approach than, maybe, we've seen technology incorporated into health care in the past?
Dr. Lozovatsky: I think so. I think what we're finding here is that the utilization of the AI tools is very targeted. And it has not necessarily been before because many of the technology implementations in the past were really aimed at digitizing records, whereas now, we're thinking about how to incorporate these tools into existing infrastructure.
And so what we found with the utilization of these AI tools is that, as you mentioned, they're really aimed at administrative burden reduction. And for over half of our physician respondents, that was the top priority—most, of course, using ambient tools but, again, not exclusively.
We also asked them some questions about what are some of the top attributes for them to adopt these tools. And what was really interesting is that response also has changed over the last year. And what we're finding now is that, for our physicians, the very top priority is having a feedback loop, which makes a lot of sense.
And it was great that it came out in the survey because what it really means is that the physicians are starting to recognize the value of these tools. They're excited about them. And they're nervous that if something goes wrong, they will not have a way to report it. So this feedback loop is really important. And of course, data privacy and thoughtful integration were close second and third priority, which, again, makes sense because the physician workflow is so critical to our ability to care for patients.
Unger: So just to clarify, that feedback loop, if it's not properly implemented or they're feeling, maybe, that something went wrong, there's a place to raise your hand and say, I need help.
Dr. Lozovatsky: Yeah, absolutely, a place to raise your hand, say I need help, and then an expectation that the issue will be fixed. I think that's also a huge piece of this that we sometimes forget, that systems need to put these processes in place so that the physicians feel comfortable, that if the tools aren't working the way that we expect that we have the ability to quickly pivot and make changes, and at the end of the day, take care of our patients. Because it's the patient outcomes that we're all concerned about with the tools.
Unger: And this just echoes—I had a conversation with another physician about AI use in an emergency setting. And what really struck me was you mentioned usability and how far it's come. But the way that she described how it was implemented, it was so much about observing and talking with physicians about the implementation. And so, right away, there's got to be an improvement in that design that, combined with the feedback loop that you're talking about, would make usability really advance.
Dr. Lozovatsky: Yeah. We've learned over the years that the needs of our clinicians are very different. So I'm a pediatric hospitalist. The needs that I have when it comes to technology when I'm caring for patients are very different than a radiologist or a pathologist or a nephrologist. And so having that conversation with the clinicians and having the ability to provide real-time feedback is going to become more and more critical, especially because these tools are moving so quickly. And we need to make sure that we are very thoughtful about how we monitor them over the long term.
Unger: Speaking of moving quickly, you mentioned the pace at which you're observing advancements. What do you think we're going to see a year from now?
Dr. Lozovatsky: It's, of course, hard to say because, as we have seen in the last year, so much has happened. But I do think that, in the next year, many of these tools will become table stakes, which is hard to believe because they're so new. And yet we're seeing now over 60% of our clinicians already using them.
I also think that we have now really recognized the fact that the technology is moving faster than the people and the process and the conversations that we just discussed that need to be had with the clinicians and the integration that has to happen. And so, over the next year, we're going to learn a lot about how to integrate these tools thoughtfully. I feel like it's going to become a huge differentiator for the organizations and the practices that do this well versus those that don't.
I also think that we're going to, over the next year, really start to have a more targeted approach. And what I mean by that is we've had a lot of excitement with ambient listening. There are people that love it. And yet we're still finding a significant subset of clinicians that it's not working for.
We also initially got very excited about the automated in-basket responses. And yet we have found over the last year that some of those aren't working as well as we anticipated, particularly for the physicians, because they're spending so much time editing. And yet some of the other clinicians are loving them.
And so as we roll these things out, I think in the next year, we're going to have a lot of learnings. And we'll be able to pivot and start to really think about the ROI of these tools and where they fit in into the clinical workflow and which of the individuals will benefit most from them.
Unger: It sounds a little bit like how I used to think about the term "digital transformation" when I was a chief digital officer and how, really, nobody knew what that meant, and it was very scary. But when you go to the specifics and the use cases about how you use digital technology to advance your objectives, you'd look back in the rear view mirror and said, hey, it was a digital transformation.
It sounds like AI is the same thing. It's a big and somewhat undefined concept. When you get down to it, you're talking about ambient listening or talking about automated responses for inbox. Those are real use cases, real sources of burden. As you think about how to expand using AI over the next year, is it better to start thinking about that?
Dr. Lozovatsky: Yeah, absolutely. I think that any technology really should be aimed at what are the problems we're trying to solve. And while we have some theories about how to use this technology to solve the problems, in many ways, it's going to be an opportunity to slow down a little bit and really understand, is the technology doing what we expect?
And much of what you described with digital transformation, this is going to be very similar, that we're going to need to think about our current workflows. How do we potentially pivot from the way that we do things because the technology can allow us to do them more efficiently and effectively in some places?
And who is the appropriate audience for these tools so that they are being utilized to the best of the ability of the technology? And this is when I say that the tech is there. It has moved way faster than we, as humans, have moved. And so, over the next year or two, it's going to be a time for us to slow down and really thoughtfully integrate.
Unger: Now, you have a really interesting vantage point to watch this work and influence it with the AMA. Can you tell us a little bit about how the AMA is helping practices navigate AI and implement it in a way that really works for physicians?
Dr. Lozovatsky: Yes, absolutely. It's been an honor and a privilege to take on this role because, as you mentioned, we at the AMA are making sure that we're thinking about the needs of our clinicians in developing tools that are most helpful to them. And so I want to mention a couple of the things that we're doing.
First of all, the AMA has rolled out AI education modules, which I think are fantastic. And the reason they're so important is not only for our trainees but for all of our practicing physicians. We have to keep up with what's happening with this technology and understand the basics. That doesn't mean that we're all going to become CMIOs or CIOs, but it does mean that we have to at least understand, when we're using the tools, how they work. And so I really encourage folks to take these introductory modules to start to learn about that background.
The other thing that we have just developed is an AI governance toolkit. And we talked about the fact that the technology is moving fast. It's going to be more important than ever to really understand how to build that governance in your practices, in your organizations.
How do you determine which of the tools to use? How do you evaluate them? What are the right skill sets for those individuals? How do organizations monitor them over time? That toolkit has all of that information.
And we're also running an AI governance collaborative, where organizations are working together to use the toolkit and think about how to do this at their institution. And we will be highlighting many of those in some of our podcasts. And finally, I want to mention that we have a new task force for AI Digital Health and Clinical Informatics, which is comprised of leaders across the country that are constantly thinking about, what are the needs of our clinicians, and how can we continue to develop tools at the AMA to support those?
Unger: Well, that's really important. And I know without that kind of governance structure, you can have probably 50 or 60 different AI initiatives happening in one organization, and there are a lot of implications for something like that. Dr. Lozovatsky, we're going to include more information about the efforts that you described in the description of this episode. And I just want to say thank you so much for joining us. We look forward to talking with you again soon.
To support the AMA's efforts to make technology work for physicians, become an AMA member at ama-assn.org/joinnow. That wraps up today's episode. 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.
