WashU Medicine works to integrate AI into MD curriculum
Focus on using AI responsibly as a tool to enhance patient care
Matt MillerSteven J. Lawrence, MD, a professor of medicine and an assistant dean for curriculum at WashU Medicine, collaborates with Quinn Payton, a fourth-year medical student, on plans to integrate artificial intelligence (AI) into the curriculum. They are part of a working group of faculty, staff and students preparing learners to use AI ethically while leveraging it as a tool to improve education, research and patient care.
As a medical student at WashU Medicine, Quinn Payton often questioned how her professional identity might evolve in an era increasingly shaped by artificial intelligence (AI). She knew she wanted to care for patients — with that, she felt confident. Rather, she wondered what that would look like in five years, 10 and beyond.
Would patients trust her medical expertise or turn to AI? Would physicians continue to diagnose disease or increasingly rely on AI? Who would be responsible for AI-related medical errors?
“AI is expanding rapidly in health care,” said Payton, who is on track to graduate in May and plans to specialize in dermatology. “Even in the few years since I started medical school, the growth has been remarkable. It’s influencing nearly every aspect of medicine, reshaping how we learn, study disease, and deliver care. It’s hard predict how different medicine will look by the time we’re practicing.”
Such concerns are valid and one of the main reasons why WashU Medicine formed a workgroup of faculty, staff and students — including Payton — to integrate AI into the school’s Gateway Curriculum, with an emphasis on AI’s ethical implications as well as its potential as a tool for improving patient care.
Indeed, advancements in AI have skyrocketed, enabling sophisticated algorithms adept at interpreting patterns in complex medical data and images with high accuracy. With the technology, patients can benefit from improved diagnosis, prognosis, earlier interventions, precision treatments and personalized health-care plans. AI can also streamline administrative tasks, allowing for increased health-care efficiency and doctors to focus more on direct care and patient relationships.
Creating a comprehensive framework for teaching AI in medical education should not be implemented hastily, said Steven J. Lawrence, MD, a professor of medicine and an assistant dean for curriculum in the Office of Education who leads the AI workgroup. “This is not just about the latest technology,” he said. “Over the decades, WashU Medicine has adapted to fast-moving technologies, but AI is evolving significantly faster and with implications for almost every aspect of our field. Our approach has been deliberate, thoughtful and gradual to ensure AI is used ethically while effectively leveraging it to improve education, research and patient care.”
Already, students can elect to study AI — as Payton did — as part of the curriculum’s EXPLORE, an immersive, longitudinal program that helps the aspiring physicians find their niche in academic medicine. Through EXPLORE, Payton and several peers conducted an institutional needs assessment on integrating AI education into the medical school curriculum, and the findings have helped inform content development. This fall, the curriculum included a class for all students on the fundamental principles of AI technology, delving into the pros and cons.
To help navigate the complexities, WashU Medicine will host Cornelius A. James, MD, one of the leading experts in AI and medical education, to present a grand rounds discussion on the topic Oct. 13 from 4-5 p.m. at the Eric P. Newman Education Center. James is also an assistant professor in the departments of internal medicine, pediatrics and learning health sciences at the University of Michigan in Ann Arbor. The university is among the nation’s first to develop its own generative AI tools for faculty, staff and students. James is involved in multiple initiatives involving AI and medical education.
The event is part of the Inaugural Goldstein Visiting Professorship in Medical Education sponsored by the Office of Education at WashU Medicine and aimed at informing faculty, trainees and students about emerging topics.
“AI has become one of the most pressing issues in medical education,” said Eva Aagaard, MD, vice chancellor for medical education, senior associate dean for education and the Carol B. and Jerome T. Loeb Professor of Medical Education. “Dr. James is truly a visionary leader in AI in medical education and in medicine in general. We are incredibly fortunate to have him visit WashU Medicine to help equip our faculty, staff and learners with the skills needed to navigate the rapidly evolving field.”
Among the ways WashU Medicine plans to teach AI is to thread it through all aspects of the curriculum. “We see AI as a valuable tool for physicians — not to replace clinicians,” Lawrence said. “This involves integrating AI tools into medical education, alongside traditional learning about anatomy or how the heart functions, to better diagnose diseases and improve health outcomes for patients.”
Other important considerations include bias and fairness. If AI systems are trained on unrepresentative or biased datasets, they can perpetuate — or even amplify — existing inequities in health care, leading to differences in treatment based on race, gender, socioeconomic status and other factors. Additionally, while AI has the potential to improve care, patients treated at hospitals or clinics without access to these technologies may not benefit equally, creating new gaps in health care.
The curriculum also aims to teach students how to become critical thinkers when using AI, said Philip R. O. Payne, PhD, a global leader in informatics and data science who is collaborating with the AI workgroup. He also co-leads the Center for Health AI, a partnership between WashU Medicine and BJC Health System to advance the design, adoption and use of artificial intelligence to enhance health care operations and the delivery of care. For example, the curriculum will teach students to appraise the credibility, impact and relevance of studies, protocols and other information that inform patient care by examining the datasets inputted into AI, the AI models used, and the questions researchers asked AI.
Similarly, students will learn how to achieve accurate results from AI by writing focused and narrowed prompts or queries. “Placing parameters around how AI, and in particular Generative AI, is used and how the results are conveyed will generate more trustworthy results,” said Payne, who is also the Janet and Bernard Becker Professor and director of the Institute for Informatics, Data Science and Biostatistics (I2DB).
Overall, Payne said, the curriculum seeks to reframe the conversation about AI from portraying it as a big, abstract technology to focusing on its practical applications in clinical care, such as assisting physicians in identifying cancerous tumors on radiology imaging that would otherwise be imperceptible the human eye, or predicting a hospitalized patient’s risk of sepsis or the likelihood of blood loss during surgery. “By telling stories of real clinical scenarios and research problems, we’re showing students how AI can help shape medical decision-making,” he said.
“AI alone should not diagnose patients and dictate treatments,” Payne added. “AI should be used as a tool to empower health-care providers and patients by improving documentation and data accuracy, and focusing their attention on crucial health signals, thus making care more affordable, safer, and of higher quality. However, the emphasis remains on the human connection between physicians and their patients.”
Anecdotally, ambient scribes — AI notetaking tools that transcribe conversations between physicians and patients with their consent — have helped strengthen such connections, said Albert M. Lai, PhD, chief research information officer and deputy director for I2DB. Used in a growing number of WashU Medicine clinics and in doctors’ offices nationwide, the “scribes” free doctors from the computer and allow them to engage more personally with patients.
The ambient scribes also allow doctors to read body language, which can often guide the questions they ask patients. “No longer is a patient talking while the doctor is focused on the computer,” said Lai, also an assistant dean for biomedical informatics and a professor of medicine and of computer science. “Our patients are pleased with the ambient AI scribes because they feel more connected with their doctors.”
For Payton, the AI workgroup has helped sharpen her understanding of how these technologies will influence her identity as a physician. “Collaborating with faculty and staff has given me a more holistic view of AI’s potential to improve diagnosis and treatment as well as engage with patients, while also being mindful of its limitations and ethical implications.”
