Members of WVSOM's AI task force

Artificial intelligence task force helps WVSOM navigate evolving technology

Imagine entering a pod-like device — think a more advanced version of the blood pressure kiosks common at pharmacies— and coming out with an artificial intelligence-powered personalized health report and treatment plan. You might sit in a chair that measures your weight, put on a jacket that tracks your vitals and be asked, “Your weight is a little high for your height; can I develop a meal plan for you?” or “Your heart rate is very high; would you like to schedule an appointment with a primary care physician?”

The chair of the West Virginia School of Osteopathic Medicine (WVSOM) AI Task Force thinks a world with this and other artificial intelligence-assisted health technologies could soon be a reality.

“I see that being not too far off,” said Machelle Linsenmeyer, Ed.D., WVSOM’s assistant vice president of institutional effectiveness and academic resources, who also participates in the AI Advisory Workgroup with the American Association of Colleges of Osteopathic Medicine (AACOM), an organization that supports all U.S. osteopathic medical schools. 

“On AACOM’s task force, we recently talked about the impact a device like this could make in rural settings where it’s hard for patients to access providers to get individualized plans,” Linsenmeyer said.

WVSOM’s own task force incorporates feedback from campus constituent groups ranging from faculty and students to departments in charge of academic technology, library resources, admissions, marketing, business, information technology and other components of the school’s operations. The group’s goal is to work toward finding ways to use AI-based tools to assist in medical education and increase work efficiency.

Artificial intelligence has been around as a concept since the 1950s and was used in primitive forms in the 1970s and ‘80s. By the 1990s, AI was helping to solve problems in technology-related fields, and as the internet facilitated access to large data pools in the 2000s, AI’s usefulness increased exponentially. It was in 2022, with the launch of ChatGPT and other large language models, that AI reached the mainstream, finding users across nearly all industries.

At WVSOM, artificial intelligence is now used as an educational tool. Some faculty members employ AI to create exam questions from course materials, and academic staff have developed “agents” — programmable systems that use AI to pursue specific goals and complete tasks — to help students create study notes and practice communication skills, among other uses.

Linsenmeyer said the creation of the school’s task force was driven by the fact that medical students everywhere were early adopters of AI-based learning tools.

“Students were already creating AI agents and developing prompts to make studying more efficient,” she said. “We had students who knew tools faculty didn’t know, and we didn’t have policies about what students could or couldn’t do with AI. The technology is moving so fast that we felt we needed to pull together information to prepare faculty for integrating this new technology into their teaching, and to bring together stakeholders across WVSOM to talk about our AI priorities.”

Barry Ndeh, of the Class of 2027, is the student representative on WVSOM’s task force. He was first exposed to AI toward the end of his time as an undergraduate and began using it as a study tool during his first year of medical school, prompting it to create lecture outlines, practice questions and flashcards.

“So much of my study time was spent simply trying to create study materials that I didn’t have time to actually study. When I started using AI to make the materials, I was able to cut maybe five hours a day from my study time, and my grades shot up,” he said.

Ndeh soon shared his agents with classmates. At first, the tools he generated didn’t get much traction, but by the start of his second year, students were taking notice — especially after he hosted an informal artificial intelligence seminar for members of WVSOM’s Class of 2028. 

In April, Linsenmeyer and Ndeh offered training as part of a pre-conference AI workshop in Arlington, Texas, at AACOM’s annual Educating Leaders conference. The audience largely consisted of deans and professors from U.S. osteopathic medical schools.

“It blew them away what you can do with it, and the efficiencies it creates,” Ndeh said. “There’s an application that lets me plug in a topic and AI will teach me based on my favorite show — for example, ectopic pregnancy using a made-up scene from Family Guy. I showed them and they said, ‘Wow. This is insane.’”

Linsenmeyer noted that WVSOM’s faculty also is learning to use artificial intelligence to evaluate students’ SOAP notes — the health care industry’s standardized “Subjective, Objective, Assessment and Plan” system for documenting patient encounters — and for assessing scholarly activity such as literature reviews.

“We’re developing more efficient ways of grading by creating tools that allow you to put these things into the system. Usually, faculty have to manually grade for 200 students. But with AI, they can just check to make sure they agree with its suggestions. And we didn’t just throw this out there; for example, we took actual scholarly activities students submitted last year and put them through the AI agent, then refined the process to get to a point where we felt grading was being done accurately,” she said.

Additionally, Linsenmeyer said the school is looking into ways to automate curriculum mapping — a process that shows how a curriculum aligns with intended learning outcomes — and is using AI-enhanced simulation cases in WVSOM’s Clinical Evaluation Center (CEC), where students hone their medical expertise through realistic, interactive models such as manikins, simulated patients and virtual reality.

Emily Thomas, D.O., medical director of the CEC, noted that the center recently changed to a virtual reality clinical simulation platform partially powered by AI.

“We made the switch because the new platform, Lumeto, uses generative AI, which allows for more realistic dialogue between students and the simulated patient, with fewer pauses and a wider range of emotional states, cadence and voice quality compared to our previous VR platform,” Thomas said. “The platform uses AI for in-goggle experiences and has added a screen-based component that will allow students to do practice simulations from home with meaningful, curated feedback.”

Other potential resources being considered by WVSOM’s administration are a Student Government Association-approved artificial intelligence club and an elective course devoted to the role of AI in medical education.

Any new technology brings challenges as well as benefits. Students’ early embrace of artificial intelligence meant that one priority of WVSOM’s AI Task Force was ensuring protected intellectual property isn’t inadvertently released broadly.

“Students didn’t realize we have faculty whose content is their property, and if they were putting faculty materials into a public large language model, the model will be trained using that information,” Linsenmeyer said. “So during this year’s orientation, we educated students to set up permissions or use WVSOM-protected resources so that what they put in won’t be distributed. We’re trying to support students in ways our faculty and administration feel comfortable with so that everyone is on the same page.”

Katie Williams, Ph.D., WVSOM’s senior director of academic technology and faculty development, said that although AI and the changes it brings can be alarming to some, its role is to supplement traditional medical education, not replace it.

“There are real concerns about the implications of AI, such as the idea that it might replace jobs in education. It’s important to remember that while AI is great with information, it can’t replace the relationships faculty and students build in learning environments. Artificial intelligence can help synthesize information, but educators are the ones who help learners take the knowledge gained through AI and transform it into real-world skills they’ll bring into clinics and hospitals. At WVSOM, we’re using AI to make educational encounters stronger by giving students tools to increase their ability to move to higher-order thinking,” Williams said.

The use of AI even extends to recruitment and admissions — not just at WVSOM, but at osteopathic and allopathic medical schools throughout the world. During the 2024-25 academic year, WVSOM’s Office of Admissions participated in an AACOM pilot program that taught staff how to use a predictive AI agent to determine which applicants are likely to matriculate at a given osteopathic school.

Ronnie Collins, WVSOM’s executive director of enrollment management, said the agent will be implemented with the next enrollment cycle as a part of the school’s screening process for students applying for the 2026-27 academic year. Admissions staff provided four years of enrollment data to teach the AI agent to look for factors common in applicants who choose WVSOM as their medical school, such as a student’s undergraduate institution, geographical distance from Lewisburg, rural interest and amount of health care experience.

“It will help us identify candidates who have different metric combinations we might not have been able to identify manually,” Collins said. “It pulls applicants to the top of the list if they meet the profile of our students based on historical data.”

Artificial intelligence processes won’t play a part in making final decisions about whether a student is invited to attend WVSOM. There are human factors machines can’t measure, Collins explained.

“The system doesn’t take into account things applicants do outside it. If a student comes for a campus tour or an open house or second-look day — or if an alumni writes them a letter of recommendation — AI won’t know that,” he said. “But we get more than 5,000 applications each year, so a funneling process has to happen with or without AI. This just makes it more efficient and gives us a first look at who’s likely to come here.”

AI also can help admissions staff pinpoint undergraduate schools to target for in-person recruitment events. Collins said the AI agent has identified Commonwealth University-Bloomsburg and Juniata College, both in central Pennsylvania, as institutions producing increasing numbers of WVSOM matriculants.

“We’ve noticed trends of people coming from these schools, so we’re sending recruiters there. We’re increasing our efforts in places we may not have looked at before,” he said.

James W. Nemitz, Ph.D., WVSOM’s president, said AI provides efficiencies that make it useful to many members of the campus community, and that without a task force to guide its use in medical education and workflows, the school would risk being outpaced.

“We have recognized that artificial intelligence provides a competitive edge whether you’re a student, faculty member, staff or administrator,” Nemitz said. “It’s here to stay, and we have a choice to either incorporate it into our work or be left behind.”