AI Has Entered the Med School Building

As it stands in 2024, there’s virtually no field immune to, or untouched by, artificial intelligence (AI). In the 2 years since OpenAI launched ChatGPT into the world, the technology has seeped into every corner of our lives, and medicine is no different. As it becomes increasingly ubiquitous, medical schools around the country are taking note and determining how AI fits into curriculum.

Right now, there are no standards regarding how medical schools weave AI into their programs, but there is some emerging commonality. Some, like Harvard Medical School, are addressing it out of the gates with first-year students. Others, like UT Health San Antonio are offering an AI dual degree program along with partner The University of Texas at San Antonio. No matter how it is presented, it’s a sure bet that AI education will be a part of med school teachings going forward.

“Every level of medical education is being impacted, and it is an exciting time to work in academic medicine supporting and amplifying these innovations,” said Lisa Howley, PhD, senior director for transforming medical education at the Association of American Medical Colleges (AAMC).

According to a 2022-2023 curriculum survey of MD- and DO-granting medical schools in the United States, 45% (61 schools) reported covering AI in required curriculum, and 47% (62 schools) reported covering AI in elective or optional curriculum.

“Artificial intelligence is impacting both healthcare and education in important ways. Medical education is unique in that it intersects both sectors,” said Howley. “Those who are responsible for designing and delivering medical school curricula must keep up with changes in healthcare delivery and practice as they prepare future physicians to practice effectively in dynamic environments.”

The American Medical Association sees AI in medical education as a necessity, with real benefits and potential.

“We believe it can enhance learning outcomes and patient care,” said Jesse M. Ehrenfeld, MD, MPH, immediate past president. “But we want to see it ethically applied and implemented transparently.”

Medical schools — while enthusiastic about AI’s potential in learning — have the same concerns and are rolling out their AI curriculum with that in mind.

“If we don’t take a fundamental role in the development of AI in medicine, the tech people will focus on the profit side of the equation and this will be a lost opportunity,” said Ronald Rodriguez, MD, PhD, professor of medical education at the Joe R. and Teresa Lozano Long School of Medicine, program director of the MD/MS in Artificial Intelligence dual degree at the University of Texas Health Science Center at San Antonio. “We need to take a leadership role and focus on patient outcomes from the beginning.”

Curriculum Development

As dean for medical education at Harvard Medical School, Bernard Chang, MD, had generative AI on his radar as soon as ChatGPT launched and began its rapid fire spread. He recognized it as revolutionary, he said, and knew the tool would soon become essential in medical education.

“I personally think GenAI represents a tech revolution that comes around every couple of decades and makes us rethink how we teach our doctors,” he said. “The last time this happened was with the Internet in the late ‘90s and early 2000s.”

Back then, the internet represented a new era for doctors, one in which they no longer needed to keep all their knowledge in their brains — they could now look things up, if needed. Today, AI is ushering in the opportunity for doctors to move beyond low-level cognitive tasks and into higher order tasks, said Chang.

“As a first-year med student, you learn how to talk to a patient, do the exam, and then record that information in a note,” he said. “After that, writing clinical notes becomes a force of habit.”

This is where AI could step in, said Chang. The tool can draft a clinical note, saving physicians time and offering them a chance to spend more time examining and talking with a patient.

“It’s harder for AI to manage the higher-level tasks, like offering the patient treatment options and understanding of their diagnosis,” he explained.

With this and other examples in mind, Harvard has partnered with the Massachusetts Institute of Technology to offer first-year students in the health sciences and technology track a 4-week class on AI in healthcare. The program launched this fall and introduced 30 students to machine learning, use of data, use of AI, and finally, the use of GenAI.

“We think it represents what we imagine to be the power of AI for doctors of the future,” explained Chang. “This cohort of students will follow it up with another 4 weeks of project-based AI work in January.”

In Texas, the approach looks a bit different in the 5-year, dual degree program. “This is a pivotal moment in healthcare,” said Rodriguez. “Our workforce is aging and we’re not producing as many doctors as we’re losing. Using AI in the right way can improve patient care as we lose manpower.”

The UT program is a multidisciplinary degree program with three tracks: Data analytics, computer science and intelligent and autonomous systems. In the AI program, students will get hands-on training in its medical applications, among other things.

“In January, we’ll be introducing students to patient-care scenarios and teach them how to use the tools,” Rodriguez explained.

At the College of Medicine at the University of Florida, the staff has developed online courses to teach medical students how AI can improve medicine. Azra Bihorac, MD, MS, senior associate dean for research affairs at the school, said, “The face of medicine will be changed with AI, and we must upskill our workforce.”

To that end, the school hopes to launch a dual degree in medicine and AI next fall. “We have an AI passport program and will launch a three-level program for our medical students,” she said. “We believe we can create a community of practice.”

Navigating the Downsides

No matter the program, everyone involved in developing medical school AI education is cognizant that they must roll it out with caution.

“There is the potential for AI to exacerbate disparities in healthcare and education,” said Chang. “We need to be intentional in avoiding that. It’s a tool and if used correctly, can combat those disparities and biases.”

Rodriguez concurs. “The potential to harm is very high,” he said. “But there are multiple groups that are aligning to minimize that.”

It’s also important that students learn the basics first, before moving on to AI, said Chang.

“That requires some care on our part,” he said. “We don’t want students to jump the gun and rely on AI as a substitute. It’s like a calculator — it’s been around for decades but we first teach math in elementary school. Once students have the basics down, we give them the tool to move on to higher math.”

The AAMC is watching the ethical aspect closely.

“The role of social, economic, and political systems on AI-based tools and how these relationships impact ethics are important,” said Howley. “An ability to remain vigilant and practice safely when technological interventions become even more standard will be critical.”