This is Ozarks at Large. I'm Kyle Kellams. Surging medical journal articles about the possible effects of AI on health care reveal a common theme: artificial intelligence will certainly have a disruptive, powerful and transformative effect on how medicine is distributed — and its full impact remains unknown. We are at the very dawn of AI's role in medicine. All of this is the inspiration for a course offered next semester by the Honors College at the University of Arkansas: AI and Integrative Health. The course will be led by Alejandro Martin-Gomez, assistant professor of electrical engineering and computer science, and James Abbas, professor of biomedical engineering. Both conduct research through the Institute for Integrative and Innovative Research, or I³R, at the University of Arkansas.
Alejandro Martin-Gomez says, clearly, AI is here to stay.
Martin-Gomez: When we start thinking about designing this course, we also thought about the relevance that this tool could have for health care, particularly for the country. We see that there are many things that need to be addressed to improve health care in general, and we think AI has the possibility to achieve that if it's used properly. That was the reason why we wanted to start talking about this.
Kellams: What sort of topics will you cover over the course of this semester?
Abbas: One aspect of this is AI — everybody's familiar with ChatGPT and large language models. But the other part is embedding AI in medical devices and health devices. We're not only looking at how ChatGPT and things like that might affect health and health care, but how will these things be embedded in devices that people using them may never realize that AI is there at all. That aspect of it is important to consider — the breadth of how AI is being used.
Kellams: Can you give an example of where AI might be embedded in something that relates to health?
Abbas: One of the main areas that I've been working in throughout my career is neurotechnology — how do we use technology to interface with the nervous system? In particular, I've been focused on how we can use that to provide assistance to people with disabilities and help overcome the challenges that they face in their daily lives. With these neurostimulation systems, there's a large number of settings that need to be in place, and if we're applying the same device for two different people, those settings are going to be different. The device has to be personalized for that individual, customized for them. Today, the way most of these are done is through a manual process, pretty iterative, with a very highly trained clinician. AI has a lot of potential to come in and say there are ways that we can do this that would be more automated. We might get better results, but we certainly will open the door for providing access to these technologies in a broader sense, so that you don't need to have somebody who has years of experience and was trained and very experienced in using these techniques. You might be able to have a less highly trained clinician being the one to deliver that technology. That's something that I personally — and I know Alejandro shares with me — is trying to figure out: how can we use technology to make it easier to deploy, so that you don't need a specialist in every case. That's one area that I think AI can really make a difference.
Kellams: Alejandro, I'm guessing that AI could be used productively throughout the health ecosystem.
Martin-Gomez: Absolutely. There are many stages where we can integrate AI into the health care system, and that's one of the things we really want to discuss. We want to talk about health care from a more holistic approach — starting from how do you eat right, how do you exercise. Try to see approaches where AI can, for example, guide you to create a meal plan, or help you how to exercise depending on the resources that you have available. We're trying to cover all the spectrum — from trying to live a healthier lifestyle, all the way to the point where you now are sick or have a disease, and how AI can help improve the workflow to make clinical interventions safer or more effective. That's kind of where my research has traditionally been.
Kellams: Can I ask a really naive question — what does your research look like?
Martin-Gomez: The core of my research is extended reality. Extended reality is an umbrella term that we use to refer to virtual reality, which a lot of people are familiar with, but it also covers other technologies like augmented and mixed reality. These are technologies that basically enrich our perception of the real world by adding computer-generated content, and that's where AI can play a very important role. I've been trying to use and integrate that technology for use in surgical procedures, mostly so surgeons can have visual information that can help them complete the procedure in a safer and more efficient way. With these technologies, we are not trying to replace the physician. We are trying to enhance their capabilities. We want to make the procedure safer and more efficient.
Kellams: This is the Honors College, so you're going to have students from across campus — some in engineering, some in arts and sciences, from everywhere. Do you think that can be beneficial for the discussions?
Abbas: Absolutely. As a biomedical engineer, it's very rare to work on projects where you're only dealing with biomedical engineers. You're working with clinicians, with scientists, with mathematicians, who all have a very different approach. But then we need to take it a step further and look at people that are thinking in terms of society — sociologists, psychologists, policy-related issues and all. The word that we use at I³R is convergence. As we're trying to address these really challenging problems in society, we need to get people from different disciplines and from different sectors. Even as we work in our research projects, we need people from academia, from industry, from the health care space, from the government, maybe from philanthropic organizations, to come together to solve these challenging problems. We like to call them wicked problems. We want to reflect that in the course — provide students with that experience of why it's important for engineers to be talking to psychologists and sociologists and policy and political science majors as we discuss these things. It's important for them to communicate with each other, to convey their messages to people outside their discipline, but also to understand the messages from people outside your discipline.
Kellams: There's a pretty big gap between people who understand AI and people who've heard of AI but don't understand it. Part of what this course does is help people become more understanding of what AI is and isn't — because there are a lot of people who hear AI and are instantly either angry or afraid.
Abbas: Agreed. And there are people on the other side that hear AI and they're excited about it without really understanding what it is. We want to bring both ends of the spectrum. The people that are so excited about it — well, there may be reasons to be a little bit cautious. And people that are afraid of it — well, it's not a magic box. There are ways of putting guardrails around these uses of AI that would probably be palatable for many people.
Martin-Gomez: I don't think we really understand yet what the capabilities of AI are. There are researchers that really understand how to create these architectures, to feed your algorithm with an input and then get the output that you're expecting. But right now — and this has been part of the conversation — there are algorithms where you don't really know what is happening inside. It's just a black box that's trained to give you exactly the output that you expect. Our level of understanding is not quite there yet in terms of explainable AI. It's a topic that is pretty hot right now, and we really need to understand what's going on inside. We're also trying to explore a lot and understand how we can use AI in different fields. That's something we'll need to be careful about and understand the risks that could be associated with just using these resources.
Kellams: This is in partnership with the Honors College and I³R — can you tell us about the institute for people who may not be familiar with it?
Abbas: I³R is a newer institute. We've been really in place for about four years now, in a new building that just opened up that many people have probably seen. The institute was formed out of a consensus of people in the region who said we want this region to grow. The university is a major part of that, but could be an even stronger driver if there was an institute focused on translation and economic development — not only how do we do good research, but how do we get the impacts of that research out into the community, and how might that impact economic development in the region. Are there new companies that can be launched or attracted to the region because of the ecosystem and the entrepreneurship and innovation? The institute was formed to focus on solving these complex, wicked problems — problems that aren't going to be solved by one person coming up with a great idea, but that really need a convergence approach and input from people from different disciplines and different sectors. The other aspect of the institute that's also important for the course is the notion of deployment, not just designing. As engineers, we love to say we design great things. But there's a lot of great things that have been designed but have never really been used widely in society. At I³R, we're really focused on getting those things out and deployed into communities so that people can benefit from these technologies. We can demonstrate something in the lab, but to get it so that it's used and having impact is a very big gap. And that's something that we really want to focus on in the institute, and one that we think AI can help with, and one that we hope the students get a better appreciation for — how do we bridge that gap.
Kellams: You're talking scale, affordability, accessibility — those are all big topics.
Martin-Gomez: And I think that's one of the big benefits of having this diversity of different student backgrounds in the course. For me, it will be a successful case if we can promote conversation — having people share their perspectives and their expertise and have this more general view of the problem or the potential solutions.
Abbas: Looking at why we do this research and why we're doing this with AI in health care: we can look worldwide and there are lots of health and health care problems. As a country, we often seem to think that we have the best health care system in the world. But when we look objectively and more broadly, on many measures we're not in the top ranks. We are in the top ranks of cost. Health care per capita in the U.S. costs twice as much, and sometimes three times as much, as in other major industrialized countries. So why are we paying twice or three times as much and not getting the same quality of health care? Can AI help us in that realm?
We look at Arkansas, and Arkansas ranks pretty lowly in U.S. states in many health measures. If you look across Arkansas, there's a wide range — in Northwest Arkansas, health metrics are pretty good. If you look at other parts of Arkansas, they're not. How can we reach those areas of Arkansas where people don't have access to the kinds of health care that we have access to in Northwest Arkansas, or that other people have access to in other states, or that other people have access to in other countries? If you're in some parts of Arkansas, your life expectancy is about 10 years less than in Northwest Arkansas. If you're born in Arkansas, you're twice as likely to die than in other states in the U.S. Why can't we address those problems? A lot of these are problems in rural health care — people in rural areas that just don't have access to things. Can AI help provide that access? With our approach of convergence, can we address these problems by reducing the need for specialists? These are the kinds of things that we're working on because we think they're important, and we'll definitely be discussing in the class — thinking about what are areas we're not thinking of. In the general realm of rural health care in the U.S., there are a lot of ways to look at how AI can address these problems.
James Abbas is professor of biomedical engineering, and Alejandro Martin-Gomez is assistant professor of electrical engineering and computer science. Together, they'll lead the University of Arkansas Honors College course AI and Integrative Health next semester. Both conduct research through the Institute for Integrative and Innovative Research, or I³R, at the University of Arkansas. Our conversation took place at the Carver Center for Public Radio.
Ozarks at Large transcripts are created on a rush deadline and edited for length and clarity. Copy editors utilize AI tools to review work. KUAF does not publish content created by AI. Please reach out to kuafinfo@uark.edu to report an issue. The audio version is the authoritative record of KUAF programming.
