The proliferation of augmented intelligence (AI), commonly referred to as “artificial intelligence,” has effectively ended the debate over whether graduate medical education (GME) physicians should use this tool. What remains now is a program to determine how best to teach health AI literacy, a skill that can improve the care doctors provide and strengthen their careers once they enter the workforce.
“The conversation a few years ago was focused on all these concerns about AI. We weren’t sure if we wanted to promote the use of AI, so we weren’t teaching about AI. Fortunately, we’ve moved past that and are now at a point where we recognize that indeed those concerns are creating a need to teach people,” said Kimberly D. Lomis, MD, AMA’s vice president of medical education innovation.
“If you’re worried that AI is amplifying bias, or if you’re worried that your population’s skills are being diminished because of the use of these tools, the only way to start the conversation is to have it transparently and provide training and foundational knowledge,” Dr. Lomis said. She helps lead the AMA ChangeMedEd initiative. The initiative continually works with forward-thinking partners to create and disseminate innovations to improve physician training to meet current and future patient needs.
The AMA uses the term “augmented intelligence” as a concept that focuses on the role of AI in supplementing human intelligence, rather than replacing it.
And AI is everywhere, including in the hands of residents and fellows.
In a 2024 AMA survey (PDF), nearly two-thirds (66%) of physicians reported using some form of AI in their practice, a significant increase from 38% of physicians who said they used it in 2023.
From implementing AI to digital health adoption to EHR ease of use, the AMA is fighting to put technology to work for physicians. This includes the recently launched AMA Center for Digital Health and AI to give physicians a powerful voice in deciding how to leverage AI and other digital tools to improve the patient and clinician experience. The AMA also recently adopted a policy in its midterm session of the House of Representatives to expand AI training across the medical education continuum.
Key questions about using health AI
As residents and fellows join other practitioners to use AI as a scribe for research and many other uses, residency programs are finding ways to incorporate this tool into their GME curriculum. According to the 2024 Work Trend Index Annual Report (PDF) published by Microsoft and LinkedIn, 75% of global knowledge workers (a group that includes physicians) use AI at work, and 78% of those who use AI at work say they use their own AI. An organization’s lack of transparency regarding the use of tools increases the risk that they will be used ineffectively and that protected health information may be compromised.
“We’re pretty early in that journey, as I think most other health systems are,” says Bruce Levy, MD, director of the Clinical Informatics Fellowship Program at Geisinger Health Sciences University in Danville, Pennsylvania. “We’re still figuring out how to use AI professionally, both as learners and as clinicians, and there’s a great discussion going on about that, and I’m sure it’s happening across the country.”
Geisinger is part of the AMA Health System Member Program, which provides enterprise solutions that provide leaders, physicians, and care teams with resources to advance the future of health care.
Dr. Levy said it is important to understand how AI will impact trainees and colleagues and shape their skills as physicians.
“How might this compromise our ability to check whether our trainees have received the appropriate training? If this tool creates notes, do we need to make sure they understand how to write notes before we allow them to use it? When is the right time to introduce it into training? How often should it be used?” he said. “These are all really important questions and will take time to understand more deeply.”
Dr. Lomis said educators also need to evaluate each task that may be removed or modified to determine whether it has a putative educational benefit and, if so, are there alternative ways to achieve that educational goal.
However, Dr. Levy said he also uses AI and believes its correct application should be part of the GME curriculum.
“I use it every day in a variety of ways, and while it’s great for some things, I don’t think it’s particularly useful for others,” he said, noting that listening to environmental sounds is an example of a task that generative AI is good at and that most doctors find useful. He said it can also be useful, for example, when searching for specific items from unwieldy documents and incorporating them into a presentation. But either individual prompts or a series of prompts aren’t very useful for complex tasks or long chats, he said.
Levy added that his experience as a programmer has taught him that code generated by AI is “more brute force and less sophisticated” than code written by humans.
“There are things you can do and things you shouldn’t do” when it comes to AI, he said. “Your car can go 160 mph. Should you drive 160 mph? Probably not.”
Health AI is a new fact in modern medicine
Physicians and trainees, like practicing physicians, want and need to use AI in the delivery of care. Your educator can help you find the best way to do this.
Residents and researchers “know how to use certain applications, but often they don’t understand the basics,” Lomis said, likening it to someone driving around a parking lot in a car for the first time. “You haven’t fully learned how to drive. You’ve learned how to operate that machine, but you still have to follow the rules of the road and the risks like texting and driving and the dangers of drunk driving.”
He said it’s important to teach new doctors how AI is developed and trained, the potential for error in the datasets on which it is trained, and how different models work. Programs should work with trainees to design digital literacy educational interventions to support residents and peers who want to use AI responsibly.
“It’s very encouraging to hear about how residents are actually using it. Very few people are using it to get immediate answers,” said Dr. Lomis, who recently spoke on the topic at the International Conference on Training and Education in Ottawa, Canada. “Most of them use it to relieve administrative tasks or as a sounding board to test ideas and develop understanding, just as practicing physicians do.”
While an important topic for all physicians, AI can be an asset when it comes to reducing burnout, especially for physicians in residency and fellowship training.
“If you can reduce tasks that are not useful in the learning part of the educational process, you can free up time for other activities, including rest,” Dr. Lomis said, adding that most residents and international students are “trying to relieve some of the burden in a very thoughtful way, not to take away the learning part.”
Dana Isaacs, MD, MPH, a second-year hematology-oncology fellow at California State, said she uses it often. New England Medical Journal‘s AI tool can help you find key past clinical trials for different types of cancer.
“Obviously, experience with patients and expertise there trumps everything,” said Dr. Isaacs, who is also president of the AMA Resident and Fellow Section. “However, we must be very careful about the limitations of AI, as it may invent false sources, fake publications, or hold completely inaccurate data.”
The AMA STEPS Forward® webinar, “AI Tools for Documentation: The Newest Member of the Care Team,” provides further insight into how ambient AI tools can leverage technology for physicians, care teams, and patients.
Also, learn more about the emerging landscape of healthcare AI at AMA. Also, learn how to apply AI to transform healthcare with the AMA ChangeMedEd® Artificial Intelligence in Healthcare Series.
