As the adoption of artificial intelligence (AI) expands in labs and clinics, integrating these tools in medical education curricula becomes essential to enhance training. Students must be taught which areas of AI technology can be relied upon and which require close human oversight.

While many discussions about AI tend to be future-facing explorations of the technology’s potential, Laurah Turner, PhD, associate dean for artificial intelligence and education informatics at the University of Cincinnati College of Medicine, asserted that AI is already omnipresent throughout the medical field.

“Last year, 48 percent of physicians were already using AI tools in clinical care, 22 percent of them were using it as a second opinion, and there are already over 1,200 FDA-cleared tools that are approved and in the market,” Dr. Turner explained. “If we’re looking [through] the lens of AI and its impact on patients, I think the fact that 100 million patients were impacted in 2025 suggests it is already the standard of care.”

Dr. Turner will lead Tuesday’s Keynote Series session at the ATS 2026 International Conference, from 8–8:45 a.m. in the Chapin Theater (Level III, OCCC West Concourse), on the opportunities and challenges of incorporating AI in medical education.

As AI continues to develop rapidly, medical schools have been put in a tough position to figure out how to incorporate the technology and teach students about it in real time. While curricula and guidelines on AI require expert-driven consultation and careful curation, these institutions must also integrate them quickly, at the risk of falling behind. According to Dr. Turner, data shows that many students already regularly interact with AI-driven tools in their studies.

“Over half of college students are using large language models, with similar numbers among medical students,” she said. “Medical students are more likely to use ChatGPT if they have questions or run into trouble than to ask their professors, consult textbooks, or review lectures. We’re already seeing that the primary learning relationship for nearly half of our learners in medical schools and residency programs is now with AI.”

These trends suggest that AI’s influence in medicine will likely continue to grow, at least for the foreseeable future, which means students must be taught to use the technology responsibly. With this in mind, Dr. Turner and Cornelius James, MD, clinical assistant professor of internal medicine, pediatrics, and learning health sciences medical school at the University of Michigan, have drafted a framework that’s under review for applying AI in medical education.

Dr. Turner explained they applied a “1/x decay” theoretical concept, typically used in complex computational systems, to distinguish “reliable zones” where AI excels, “zones of inflection” where human oversight is needed, and “human zones” where AI should be ignored.

“Many people are concerned about AI getting something in the domain of medical knowledge wrong, but we’re seeing that’s actually not the case. Usually, it does great with knowledge recall, so that’s not an area we need to be too concerned about,” said Dr. Turner. “Where we see the potential for severe patient harm and error is through omission when the AI fails to recommend something critical. Physicians who have been practicing can look and say, ‘Something is missing here,’ but what does that mean for future learners who are going to be learning with AI? Are they going to be able to detect when something is missing?” 

While AI can enhance students’ learning as a useful training tool, Dr. Turner argued that it is not a substitute for human-driven medical education.

“There’s a difference between training and education. Training equips us for what we can predict, such as high-frequency patterns, routine-based processes, or memorizing drug doses. Training is where AI excels,” she explained. “Education prepares us for what we cannot predict. It’s about developing the internal capacity to reason through novelty, to adapt, and to make wise decisions under uncertain circumstances. This is where humans excel. If we forget how to educate, cultivate judgment, and adapt to reasoning, then we’ve lost the only thing that humans currently do better than AI, and I would predict will continue to do better than AI.”