How Artificial Intelligence is Changing the Future of Medicine

Find out how artificial intelligence is being used in medicine and what the future of medicine could look like with this technology.
Illustration by James Wang

When it comes to artificial intelligence, Dr. Cornelius James thinks media and pop culture too often portray it as something dangerous that shouldn’t be trusted.

“One of the challenges is that people go to extremes,” says James, a primary care physician and clinical assistant professor at Michigan Medicine. He says that people think of The Terminator and other villainous portrayals.

In reality, artificial intelligence is much less harmful than we imagine — in fact, it’s quite helpful, particularly in medicine. The hard part is convincing others that’s the case.

Tackling this very issue, James co-wrote a paper titled “Preparing Clinicians for a Clinical World Influenced by Artificial Intelligence.” The AI manifesto focuses on how important education is in integrating AI into the medical field. Clinicians, he claims, need to be educated on how artificial intelligence can offer assistance rather than be a hindrance. As he and his co-authors write in the paper, “Over time, it is likely that every medical specialty will be influenced by AI, and some will be transformed.”

So what does that actually look like in a clinical setting, and what progress are we making to bring about a future transformed by AI?

Instead of humanoid robots, AI is more akin to advanced programming that can not only compute data but also understand and interpret it. AI employs a process called machine learning that allows programs and algorithms to adapt and learn without direct instructions.

This process can take medical data and apply it to practical uses in a clinical setting. As far as current progress goes, you don’t need to go far to find it. Research being conducted right in Detroit is influencing the future of AI in medicine.

Take, for example, Dr. Phillip Levy, associate vice president of translational science at Wayne State University. Levy works on developing AI models that can analyze patterns in data sets to predict risk factors in a given population, suggesting when a person may be at risk for heart disease, cancer, or other illnesses.

“What you’re trying to predict is who, at 30, will have a stroke at 60,” Levy says.

Levy isn’t the only one using AI models to assess and solve medical problems. Dongxiao Zhu is the founding director of the Wayne Artificial Intelligence Research Initiative and an associate professor at Wayne State University. His research in the field of machine learning and applications has involved creating models to predict the occurrence of specific types of cancer and even the outcomes of COVID-19 in children. Zhu believes the next challenge moving forward is for machines to process medical images and analyze more complex information.

AI can analyze aspects of numerical data, but it doesn’t often have a human resemblance like we see in the movies. Researchers, however, are looking to use AI to make conversational bots.

Douglas Zytko is an assistant professor in the computer science and engineering department at Oakland University and the director of the Oakland HCI Lab, which does research on human-computer interaction. Currently, Zytko and his team are working on projects that assess the role of technology in today’s online dating culture and use AI to help mitigate future sexual violence.

The first phase of this project is collecting data in the form of personal relationship experiences. In order to get a large enough data sample, they are creating an AI bot that will conduct interviews with subjects who share their experiences. The big challenge is to give the bot a level of sensitivity (dare we say humanity?) when asking personal questions and responding to people’s experiences. Finding that balance is difficult when dealing with a computer, but it is crucial, Zytko says.

Researchers like Zytko are working to figure out “how we can refine AI without retraumatizing people,” he says. The key, he explains, is to develop risk-sensitive technology. Zytko plans to use this data to further understand how people use dating apps and how the apps create and perpetuate perceptions of sexual consent — or lack thereof.

AI may take on certain tasks historically performed by humans, but James believes there will always be a need for real people in medicine. Artificial intelligence can make a medical recommendation, but at the end of the day, the clinician must make the final call in most cases. In this way, a human-technology relationship will form that makes AI a part of the team. James explains this kind of interaction more as augmented intelligence than artificial intelligence. He says the way we use AI has the potential to enhance our capabilities rather than replace them.

There are those who remain skeptical about becoming buddy-buddy with artificial intelligence, but James says that this will change as educational programs help clinicians interpret AI and learn to trust it. The goal is that once clinicians develop a level of comfort with AI, trust will begin to trickle down to the patient level. As patients see the successful implementation of AI in the clinical environment, he hopes they will become more accustomed to it, understandably question it, and, eventually, accept it.

The medical field is beginning to integrate a new stage of technology, between models that assess risks, bots that collect data, and other AI that can give preliminary diagnoses, communicate between physicians and patients, and transcribe medical documents. Current research has a long way to go before we see widespread use of AI in a clinical setting, since it takes time to assess accuracy, reproducibility, effectiveness, and practicality. Levy says the amount of time depends on the type of technology, but it can take up to a decade to go from a lab to practical use.

“Ultimately, physicians can expect that we will do things differently,” James says.

Every part of the medical field will work with AI in some way, but exactly how that will look is still unclear. Thanks to research, we get a glimpse into the future of AI, but there is still so much we can’t predict. AI may revolutionize the medical field in the near future, but there are also developments 20, 30, or 40 years away that we haven’t imagined in our wildest dreams.

This story is from the Moving Medicine Forward feature in the October 2022 issue of Hour Detroit magazine. Read more in our digital edition.