View Larger Image
Kimberly Lomis, M.D. (right), vice president for medical education innovations at the American Medical Association, speaks during a panel with Alex London, Ph.D. (center) and Wendy Ward, Ph.D. Image by Benjamin Waldrum

UAMS AI Conference Expands, Focuses on Increasing Role in Health Care

Oct. 22, 2025 | As the role of artificial intelligence (AI) continues to expand in society, its role in health care is also growing. This year’s Innovations in Artificial Intelligence Conference at the University of Arkansas for Medical Sciences (UAMS) explained AI in plain language and used clear examples for a balanced look at the breakthroughs it enables and possible pitfalls to avoid.

More than 200 people attended the conference, held Oct. 7-8 on the 12^th floor of the UAMS Jackson T. Stephens Spine & Neurosciences Institute, while a virtual audience of nearly 100 watched a livestream of the event. It was jointly presented by the UAMS Department of Academic Affairs, the UAMS Institute for Digital Health & Innovation, all UAMS colleges and its graduate school, and many other collaborative community partners.

“At UAMS, AI isn’t replacing the human touch — it’s amplifying it,” said Wendy Ward, Ph.D., a professor in the UAMS College of Medicine and associate provost for faculty in the Department of Academic Affairs. “Our second annual conference showed how clinicians, educators and researchers can turn responsible AI into better access, safer care and smarter training for Arkansas and beyond.”

Last year’s highly successful inaugural conference showed that there was a deep and growing need to address AI in health care, Ward said. With that in mind, this year’s conference doubled in size, with a preconference workshop Oct. 7 and a longer conference day Oct. 8. There were also several UAMS speakers in addition to other industry experts.

Fred Prior, Ph.D., distinguished professor and chair of the UAMS Department of Biomedical Informatics, kicked off the preworkshop with a discussion of what AI can and can’t do. He cautioned those in attendance to avoid “uncritical acceptance” of the responses that AI generates, including taking potential bias into account, like any source of information.

“You’re the expert, not the AI,” Prior said.

Understanding what goes into a given AI is key to using it effectively, added Mathias Brochhausen, Ph.D., professor and vice chair of the Department of Biomedical Informatics. That means starting with good data and making it easily identifiable and understandable for a machine. Essentially, AI has to be taught how to think so that it can interpret and reason properly.

He related an anecdote from a Terry Pratchett novel in which a witch doesn’t use her magic to clean the floor and instead does it by hand. When asked why, she replies that the last person who tried to clean with magic removed the floor as well as the dirt.

“We need to tell computers: this is what is the dirt, and this is what is the floor — this is the data that I want, and this is the data that I don’t want,” Brochhausen said.

Joseph Sanford, M.D., chief clinical informatics officer for UAMS and director of the UAMS Institute for Digital Health & Innovation, talked about how AI is an essential tool at UAMS, supporting multiple aspects of the health care system, largely to help streamline more technical, time-consuming aspects of reporting and record-keeping. It is built into patient access, patient care, discharge processes and billing, among others.

“We are using AI literally everywhere,” Sanford said.

AI works best when it’s answering a specific request about something that it already knows, Sanford said, which makes it useful as an assist with large volumes of health care data. However, like any computer, AI needs regular updates to be effective. Since those updates can be slow to implement, there always needs to be a human element “in the loop” so things run smoothly.

They [AI] are not human replacement tools,” Sanford said. “They are in many ways more complicated and more expensive and more fragile at scale than a human would be.”

Stephanie Gardner, Pharm.D., Ed.D., UAMS provost and chief strategy officer, welcomed attendees on the second day and said it was important to continue learning and growing with AI so that it “enhances, not replaces” the human connection to medicine.

“Together, you will all help us think about and discuss how to best integrate artificial intelligence into our missions in a way that improves care, advances learning and fuels discovery, while keeping our focus where it belongs: on patients, communities and the human side of health,” Gardner said.

Patricia Dykes, Ph.D., RN, a professor of medicine at Harvard Medical School and research director at Brigham and Women’s Hospital in Boston, presented virtually and shared several digital innovations that Harvard had developed over the years to help patients, including Fall TIPS, a fall prevention tool now in more than 500 hospitals nationwide.

Alex London, Ph.D., the K&L Gates Professor of Ethics and Computational Technologies at Carnegie Mellon University and the author of “For the Common Good: Philosophical Foundations of Research Ethics,” spoke about the challenges and uncertainty of AI in health care. Ethical concerns remain a hot-button issue with AI, and the UAMS audience had many questions.

He was followed by Kimberly Lomis, M.D., vice president for medical education innovations at the American Medical Association (AMA), who spoke about the impact of AI in medical education. She cited an AMA study that found about 70% of polled physicians believe AI will offer an advantage in health care and emphasized that educators must start learning and adapting to AI.

“Historically, we have identified a successful learner as someone who can hold a great deal of knowledge,” Lomis said. “That is no longer useful. You’ve got [knowledge] in the palm of your hand. What we really need is someone who can question that knowledge. The most successful learner in the future will be the one who asks the most probing questions, is skeptical and critically appraises what comes out and feels a sense of duty to do that with these AI tools — as opposed to AI complacency, where they’re just looking for a quick answer.”

A series of small presentations followed by UAMS researchers on ways to use AI to analyze hospital data, improve skill development and facilitate interprofessional education. Afterward, Ward hosted a Q&A session with London and Lomis that included multiple questions from the audience on various topics, including the long-term environmental impacts of AI.

The conference closed with a poster presentation session in the foyer with roughly a dozen UAMS studies on the use of AI. Attendees and speakers alike moved from poster to poster and engaged with presenters on their research.

Corey Hayes, Pharm.D., Ph.D., MPH, an associate professor in the UAMS College of Pharmacy’s Department of Pharmacy Practice, attended both days of the conference. He said he is already using AI in some of his research to help predict early dropouts from treatment for opioid use disorder. He said the hope is that better predictions can prevent future overdoses.

“The AI world is rapidly evolving, presenting both exciting opportunities and unique challenges,” Hayes said. “I’m grateful to UAMS for hosting this conference so that we can continue to provide the best, evidence-based care for Arkansans.”

Ward echoed that sentiment, adding that AI is now another tool in the health care toolkit.

“We’ve moved from curiosity to capability,” she said.