Yale Spotlights Med Ed Innovation and Research

Over 200 faculty, alumni, fellows, residents, students, and staff from Yale Schools of Medicine, Nursing, and Public Health gathered on the medical school campus on June 5, 2025, for the 13^th annual Medical Education Day at Yale (Med Ed Day).

In her welcoming remarks, Jean and David W. Wallace Dean and C.N.H. Long Professor of Internal Medicine Nancy J. Brown, MD, described the day, which the Yale School of Medicine (YSM) Center for Medical Education (Center) hosts, as an event that spotlights excellence and innovation in medical education. Brown also discussed the growing number of programs that the Center offers to the YSM community, such as the Master of Health Science degree Med Ed Track, the new Educator Scholar Fellowship (ESF) Program, and the new longitudinal Medical Education Concentration for students, currently enrolling 48 MD students. These programs reflect how medical education innovation and scholarship extend across the continuum of learning at YSM.

The Med Ed Day workshops and oral presentations on topics ranging from “Enhancing Student Engagement in Small Group Learning—Lessons from the High-Engagement Pre-Clerkship Pilot,” to “Asian American Women in Academic Medicine Leadership: A Qualitative Study of Medical Residents' Perceptions of Facilitators and Barriers to Leadership,” also reflect this span, with presenters including faculty, residents, and students.

Similarly, this breadth of engagement was evident in the poster session, with over 90 submissions from faculty, fellows, residents, and students—a record for Med Ed Day—focused on either innovation in education or medical education research. For the first time, there was a special awards category for students because of the record number of student entries—29. See the award winners here.

The day included a graduation ceremony recognizing the 15 faculty members who completed the 10-month ESF. Each participant worked on an original project designed to enhance their department’s or the school’s education programs, such as teaching, curriculum design, or evaluation, and created a poster, which was part of the poster session. The ceremony celebrated not only the graduates, but their mentors, who supported them through the process.

A highlight of the day was the fascinating keynote address by Marc M. Triola, MD, entitled “From Bytes to Bedside: Exploring the Impact of Artificial Intelligence on Medicine and Medical Education.” Associate Dean for Teaching and Learning Janet Hafler, EdD—director of the Center—introduced Triola, who in September will become senior associate dean for medical education at NYU Grossman School of Medicine. Triola currently is associate dean for educational informatics; director, Institute for Innovations in Medical Education; and professor, Department of Medicine NYU Langone Health.

A fire alarm as Triola was about to start created some excitement, including the arrival of the fire department, requiring a venue change. But Triola did not miss a beat after relocating, as he eloquently and energetically shared his self-described “extremely optimistic and extremely positive” thoughts about how the rapid advances in and accessibility of AI are transforming health care and medical education, as well as his ideas about what is on the horizon. “Health care is the best place for AI in all of humanity,” he stated, explaining it is no longer a question of if we should use AI. Rather, Triola believes it is harmful not to use it and that we should use it more, pointing, for example, to our inability to meet current health care demands.

In framing his response to the question of how good AI is at being a diagnostician, a patient communicator, and a scribe, Triola discussed the “remarkable” and “exponential” improvements in the capabilities of AI, including significant advances as recently as the last three months. He shared that AI recently passed the “Turing Test,” where participants had five-minute blinded conversations simultaneously with a human and AI, and 73% of the time, GPT-4.5 was judged to be the human. This, he believes, has great potential for health care, and companies already have products using this technology.

Triola stated that AI now can reason and research at the PhD level. He described research demonstrating that under controlled circumstances, when all the necessary information is packaged together, AI can diagnose better than physicians, with AI diagnostic accuracy improving by more than 15% in just the past few months. Triola pointed out that information often is not neatly packaged, giving the example of having to ask a patient’s relative for the patient’s medication list. AI alone also performs better than doctors using AI, because of anchoring biases and because doctors are not leveraging AI’s capabilities—they are using it like UpToDate® to ask specific questions, and not like a collaborator to help analyze case histories.

Triola also shared research showing that patients preferred AI text responses to physicians’ responses, viewing AI as more empathetic. He provided an example of a patient traveling to Costa Rica asking a question about a medication. The physician provided a brief, informative response; AI shared similar information, but then added, “Costa Rica is beautiful this time of year. Hope you have a wonderful time,” which was perceived as a warmer response. Triola noted the criticism that flowery language is not always appropriate, for example when discussing a serious illness.

Triola shared how companies have developed AI platforms that can engage in spoken conversation. For example, Hippocratic AI has 50 AI nurses, who provide clinical advice and recruit people for clinical trials. This dramatically expands capacity—an AI nurse can talk to more than one person at a time, and there is no limit on their time, for example, he said, spending 60 minutes responding to a patient’s pre-colonoscopy questions. This has the potential to dramatically lower prices.

AI as a scribe, through ambient conversations recorded on a mobile phone app, also has significant benefits, including removing the distraction of a physician having to type while talking to a patient. Studies show, Triola said, that while AI is not a perfect solution, for example needing significant editing, 46 clinicians from 17 different medical specialties​ spend 30% less after-hours work per workday on notetaking using AI tools. NYU has embraced this technology, he said, with 1000 faculty using it in in-and out-patient clinics.​ Triola envisions a future where the phone not only records the conversation, but can add to it, for example, reminding a physician to ask a particular question.

In the second half of his talk, Triola pivoted to on how AI impacts medical education—both how it can improve training, and how it changes what we must teach. In another example of how AI is rapidly improving, Triola said that in 2021, AI responded to 45.1% of USMLE (United States Medical Licensing Examination)-style questions correctly. With the newest Open AI models, that has risen to 97% in 2025. Triola believes the 3% of incorrect responses likely reflect poor questions, since it is essentially an open book exam for AI, with access to all human knowledge and trained on thousands of USMLE exams.

Triola emphasized that students arrive at medical school with deeply established patterns of using AI, which he views as positive. At school, students are using AI systems as soon as they are available from the explosion of start-ups and companies, for example to create USMLE questions and flash cards, and as virtual patients to prepare for the OSCE (Observed Structured Clinical Examination).

Triola described NYU’s use of Precision Medical Education (PME), which he defined as, “A systematic approach that integrates longitudinal data and analytics to drive precise educational interventions that address each individual learner’s needs and goals in a continuous, timely, and cyclical fashion, ultimately improving meaningful educational, clinical, or system outcomes.” Every two weeks, AI provides a snapshot for the student and their coach and advisor, on how the student is doing and strategies for addressing any areas where they are struggling. It also provides nudges to trainees on helpful steps they can take. The goal, he said, is not to replace human coaches and advisors, but to bolster them, providing more data and digestible insights about that data.

At NYU, Triola said, AI and Generative AI are now a required part of the core curriculum. AI’s transformation of medicine will require transformations of curricular content.

Triola closed by sharing his thoughts on what is on the horizon, such as cameras interpreting what we see and humanoid robots in hospitals. Though this comes with much uncertainty, Triola thinks this is the best time to be in medicine.