Alumni Spotlight: Jonathan Chen Prescribes Lessons in AI to Medical Doctors

With medical training and technical experience — and a little bit of magic — UC Irvine alumnus Jonathan Chen, MD, PhD, aims to help doctors leverage the power of AI.

On March 17, 2025, an ABC7 newscast featured Jonathan Chen sharing his surprise at the findings of two recent studies he conducted at Stanford University, comparing chatbots to human doctors. The first study looked at accuracy in patient diagnosis, while the second study dug deeper into care-management decisions.

“I have, for years, restated the fundamental theorem of biomedical informatics that human + computer will outperform either alone,” says Chen, an assistant professor of medicine and of biomedical data science at Stanford. “It was thus startling when we did a series of studies on complex medical diagnostic and management reasoning problems and found that the computer — GPT-4 in this case — performed as well as or better than human physicians, including humans who had access to GPT-4.”

Chen, who earned his MD from the UC Irvine School of Medicine as well as his PhD in computer science from the Donald Bren School of Information and Computer Sciences, strives for a world where AI enhances doctors’ care. His findings serve as a cautionary tale to those who dismiss the arrival of disruptive technology.

“It begs uncomfortable questions,” he says, “about what the point of the human-in-the-loop is in the process if they are not substantially helping, and maybe even hindering, what the computer could do on its own.”

Combining Medicine & Technology

It’s no surprise that Chen is helping lead the charge in exploring how humans can leverage AI to improve healthcare, given his unique background.

Chen started college at the age of 13, attending the California State University of Los Angeles and then transferring to UCLA. “I started college unusually young, and my parents wanted me to go to medical school. That was fine by me when I hadn’t thought of anything more seriously,” he says. “As I pursued pre-medical studies more earnestly, I realized I was much more of a computer nerd.”

He switched from pre-med to cybernetics and, after graduating with a specialization in computer studies, took a well-paying software engineering job. “I was later laid off, along with a third of the company, amidst the 2000 dot-com bubble and burst,” he says. “I got another tech job, but I could foresee these were jobs, and not necessarily a fulfilling long-term career.” So, he decided to apply for a joint MD/PhD program with the goal of working on interesting problems in computer science that were directly related to human health.

“I’m not sure this actually made much sense 20 years ago, but UC Irvine was one of the few campuses with forward thinking on the potential of combining them,” he says. “Now, with the progressive rise of artificial intelligence applications in healthcare, this combination seems inevitable and obvious.”

Understanding Disruptive Technology

“I’ve been working on clinical-decision support computer systems that can aid medical reasoning for years,” says Chen, who leads a group of researchers in the HealthRex Lab. “As capabilities in large language model ‘chatbot’ AI systems advanced, allowing emerging systems to tackle more complicated open-ended questions, I realized I needed to rework half my research program.”

Chen saw a clear need to empirically evaluate — and to direct the safe and effective integration of — humans and computers working together. “More important than how well a computer answers a question was the more relevant question of what happens when we give these tools to humans,” he says. This question led to his studies of chatbots and human doctors.

“I believe what the findings really speak to is the arrival of a disruptive technology,” he says, comparing it to when we first gained access to the internet. “Many people simply don’t know how to effectively use emerging technologies that can not only facilitate many administrative paperwork tasks but also support complex medical reasoning, and even aid in nuanced, empathetic and purposeful high-stakes conversations.”

So how might Chen guide doctors in having “purposeful” conversations with technology?

Working His Magic

“Sometimes magic is just someone spending more time on something than anyone else might reasonably expect.” This quote, attributed to Raymond Joseph Teller of the famous Penn & Teller magic duo, struck a chord with Chen.

When he realized how much time he spent playing computer games, he took Teller’s sentiment to heart and decided to reconsider his hobbies. “After my Steam gaming account told me I spent 200 hours playing XCOM 2 and many other games, it occurred to me that if I redirected even half of that time, I could probably get pretty good at other fun but more productive hobbies.”

He decided to reconnect with his childhood activities of playing the piano and learning magic tricks, finding that they provided both a creative outlet and source of encouragement. “Learning a Beethoven piano sonata seemed impossible,” he says, “yet making very small progress consistently every day was enough to make remarkable achievements over months.” He similarly devoted his time to perfecting his magic tricks, particularly during the pandemic, and eventually became an award-winning magician.

“The daily work of advancing science is often a soul-crushing grind,” he says. “The fact that we can advance the boundaries of knowledge and state-of-the-art capabilities at all is incredible, but most people highlight the 2% inspiration in the press release without mentioning that it only comes after 98% perspiration.”

Healthcare is similarly labor intensive. “The daily work of a clinician is often a grind through administrative paperwork to make actual patient care happen,” he says.

When the toil becomes overwhelming, Chen turns to his hobbies, exercising novel skills in connecting with others.

“Magic has given me an artifact to engage with people in different ways,” he says, “and has also allowed me to hone relevant skills in presentation, communication and the interpretation and direction of human behavior that is relevant in science, education and medicine.”

Blending his unique expertise as a technologist, medical doctor, musician and magician, Chen is working to realize the synergistic effect of humans and AI, maintaining that fundamental theorem of biomedical informatics.