Preparing for a Paradigm Shift: The Inaugural UCSF AI & Education Symposium

The inaugural UCSF AI & Education Symposium convened over 350 UCSF faculty, staff, trainees, and learners to explore the transformative role of artificial intelligence (AI) in health professions education. Sponsored by a gift from the Baum family, the symposium provided a forum for collaboration and hands-on engagement with cutting-edge AI technologies.

In his keynote address, Robert Wachter, MD, Chair of the UCSF Department of Medicine, framed AI’s impact through a line from the novel The Sun Also Rises:

“‘How did you go bankrupt?’ ‘Two ways: Gradually, then suddenly.’”

AI was first described in the 1950s, but until the 1990s, decade-long “AI winters” had halted interest, funding, and developments in AI – largely due to its “perceived limitations” and “excessive costs.” But by the 21st century, advances in AI’s ability to suggest diagnoses or aid clinical decision-making had snowballed – setting off an avalanche of renewed interest in AI that disrupted those placid winters.

Gradually, then suddenly.

Collaboration and Vigilance in AI

In the face of this sudden disruption, Sandrijn van Schaik, MD, PhD, Baum Family Presidential Chair for Experiential Learning, emphasized the need for cross-campus coordination.

“Given UCSF’s talent and forward-thinking community, we’re seeing ‘a thousand flowers bloom’ in AI integration,” she said. “We need to ensure they bloom in collaboration.”

Thus, the symposium aimed to foster a shared understanding of AI’s evolving role in health professions education, convening both UCSF and guest experts to weigh in.

One such guest expert was Laurah Turner, PhD, Associate Dean of AI and Educational Informatics at the University of Cincinnati College of Medicine. In her keynote, Dr. Turner discussed AI’s implications for healthcare training, advocating for AI systems that enhance human capabilities while preserving essential human elements to meet the needs of learners and patients.

“I look toward a future where AI is not just a tool, but a partner, a collaborator,” said Dr. Turner. “It can be a force that doesn’t replace human expertise but enhances it.”

Meanwhile, Dr. Wachter urged attendees to move past traditional concerns like AI’s “Black Box” problem (the opacity of machine learning algorithms) to consider new and emerging questions.

“We don’t know the exact mechanism of action of Tylenol or of most anesthetics, but they are empirically proven to work, and we still use them,” he said. “The new problem is human vigilance and oversight. How much do we keep humans in the loop and have them be effective as safety bulwarks?”

Learner Perspectives: Becoming AI Co-Pilots

This was one question Anna Park (MS4) had in mind while attending the symposium. Hoping to conduct background research for her own scholarly project – an exploration of AI in medicine through creative writing – Park noted that many of the symposium’s predictions about AI were reminiscent of already-existing science fiction. The concern about human deskilling due to AI, for example, reminded Park of the film Wall-E, “in which humans… have all but forgotten how to perform tasks of daily living.”

The challenge in medicine, then, is ensuring that we remain co-pilots of AI – not its passengers.

In helping organize the symposium, Laura Chen (MS1), a former software engineer, co-led lunchtime discussions on AI literacy in medical training. From students to attendings, researchers to trainees, the diversity of participants engaging in these conversations was striking to Chen.

“That diversity reminded me of the broad impact AI will continue to have,” she said, “and the diverse collaborations that will be needed to tackle AI challenges.”