Discussions about artificial intelligence (AI) are pervasive. While AI is already helpful in the day-to-day, there are legitimate concerns about the perils or potential perils of artificial intelligence. In this episode of Health Connections, Dr. Carol Meyers, a professor emeritus in the University of Tennessee College of Nursing, talks with Dr. Tammy Wyatt, the associate dean for research at the University of Tennessee College of Nursing.
WUOT’s Carole Meyer: How would you define artificial intelligence?
Tammy Wyatt: To me, the most basic of a definition is a machine's ability to recognize patterns and make decisions or judge the information that is inputted into the machine.
Okay, so that's a general definition. Let's segue now to a description of AI applications in health care specifically.
I think to understand AI and health care, we have to know that there are four types of artificial intelligence. And today, we actually only have capacity for two types. And those are the more fundamental types, and I'll explain those really quickly.
The first one is called reactive machines. If I put in information, and I put in the same information, I'm going to get the same output every time in health care. An example of that would be if I were using a system, a decision support tool, something on my handheld or something on the computer, I put in information about someone's lab values, someone's symptoms, I'm going to get the same information every time until I feed it more information or different information. So that's reactive machines, or reactive AI.
The second type that we now have is called limited memory. And with limited memory, this means that the computer machines are able to learn from the data. The more data that is put into the system, the more it learns. And so we also call that machine learning. When we talk about artificial intelligence, unless you're living in the AI world, then really what we're talking about are those two types of artificial intelligence.
There are two other types of artificial intelligence that are not here yet. But we are seeking and moving forward to those two types. One is called theory of mind. And what this type of artificial intelligence means is that currently today, our AI does not have the capacity to understand emotions and thought, when we reach the capacity and AI of theory of mind, then the machines will be able to understand emotion and thought, the fourth and final, the Gestalt, the guru of the, the big kahuna of artificial intelligence will be self awareness. So when AI is able to be aware of itself, at that time, it will understand how their emotions and their thoughts are not only impacting their functions, but others functions. Those are in the future. We're not there.
What I've heard you describe is quite different than if I was out on the street and put a microphone in someone's face and said, What do you think AI is in healthcare? Could you respond to that?
In preparation for our meeting today, I really wanted to do a quick informal survey of exactly what kind of answers I would get if I asked that. And I wasn't really surprised. I think when most people think about artificial intelligence in healthcare, they think immediately robots and artificial intelligence is not a robot. A robot is hardware. And what makes a robot function is the artificial intelligence within science. So when we really talk about artificial intelligence, and health care, it is the behind the scenes, functions and workings that helped to inform the care that is being delivered. It's also one of the things that artificial intelligence is not is it's not care. It's not human care. And so, when people often talk about artificial intelligence and fears about it, it's that it will replace humans.
Why is AI getting so much traction in health care?
I think the reason that AI is such a good match with health care is because if you reflect back on what how we just defined artificial intelligence, one of its the current capacity of AI is that it can learn from itself. It can learn from data.
And there is quite possibly no other industry that has more data than health care. And so, if we have a machine that learns from data, and we have an industry like healthcare that has volumes of data, in fact, we often call it Big Data than the to go hand in hand. And so not only have we been already using artificial intelligence and healthcare, but we're going to see a lot more of it in healthcare.
Let's do that. Let's bring it down to a level that would be commonly understood. There's all this big data. And it's very exciting, what some of the promises, what are some of the applications, so you know, where you might predict that AI would have a role in health care.
I think we're going to see more and more of what we're already doing. And here are some of the things that are already happening. There are already artificial intelligence applications that are able to interpret radiographs, or any kind of imaging, not all imaging from healthcare images, but some images. And while today, we're not relying solely on those, it's that the artificial intelligence is giving their interpretation. And then a human person, a radiologist, a trained radiologist, is either denying or, or validating the findings. That's one of the things that's happening now. And I anticipate that that is only going to grow and grow across other sectors of healthcare.
For example, we've had this combustion of, of data with genetics. So now that we have the ability, we have the machines that can handle the data, and we have the data, we now can give care based on somebody's genetics and how they may respond to the care that's really quite different than what we have been doing in the past, which was a standard of care. That means that if you as a female that's close to the same age that I am, if we walked into the ER and we had a heart attack, there would be some very standard of care that we would practice just right off the bat, because there's no other way that we know to do it other than the standard of care. But in the future, it might not look like that. It might be that the treatment for my heart attack, or my myocardial infarction is going to look very different than what it'd be for you. And I think that's an example of how artificial intelligence is going to really expand.
What are some of the issues that we have to deal with in terms of the perils associated with AI and health care?
I think one of the major concerns that we have is ethics. We have concerns about who is overseeing how artificial intelligence is going to transform our everyday life -- who's controlling that? How are we going to keep the demons out and make this good for the whole? And the only way that's going to happen is if we have an interdisciplinary team of individuals that are advancing AI, including ethicists, including patients, including caregivers, including health-care providers, including AI developers. It has to be a very robust interdisciplinary team that's moving artificial intelligence forward.
Do you have any parting words for our listeners?
I feel very enthusiastic about the future with artificial intelligence. In fact, it took some digging for me to really be able to align with the perils of artificial intelligence. I think the future is bright and we have a lot to look forward to, and less to be fearful of, as long as we are all engaged in the conversation, and we are pursuing it with good intentions and ensuring that the individuals that are at the table making decisions represent every single person that's affected by artificial intelligence.
This transcript has been lightly edited for content.
