{ LIFE SCIENCES }

Leveraging Artificial Intelligence to Improve Health Outcomes

The newly established Dunleavy Fund for Clinical AI helps position Harvard Medical School to lead a revolutionary shift in medical education and health care

a stethoscope on a microchip

Leveraging Artificial Intelligence to Improve Health Outcomes

{ LIFE SCIENCES }

Incorporating AI into health care promises substantial benefits like better patient care and innovative research, but how best to integrate this technology remains an open-ended question. How do we provide computer scientists and engineers with the tools to develop safe, effective AI models for clinical care? How do we cultivate AI experts who understand the complexities of medical practice? How do we use AI to accelerate diagnoses and personalized therapies?

Inspired to tackle these questions, Harvard Medical School (HMS) alums 
Keith Dunleavy MD ’95, founder of health care technology company Inovalon, and 
Katherine Dunleavy MD ’95, a physician specializing in internal medicine, established the Dunleavy Fund for Clinical AI at HMS. Supported by a gift from the Dunleavy Foundation, the new fund will strengthen the School’s efforts to equip future scientists with the interdisciplinary knowledge and skills to harness AI in service of improved patient care and medical science.

“By supporting training that brings these fields together, we hope to help in some small way to bring the power of AI to the great needs of medicine and health care,” says Keith, who is a member of the HMS Board of Fellows.

Katherine, a member of the HMS Advisory Council on Education, hopes that by emphasizing education and training in AI that is specifically geared toward medicine and health care, the fund will facilitate the growth of well-rounded professionals.

“We hope this approach brings an important element of being mission-driven with respect to the cornerstones of medicine and health care: those based on caring, empathy, and positive impact for society,” she says.

Keith Dunleavy MD '95 and Katherine Dunleavy MD '95
KEITH MD ’95 AND KATHERINE MD ’95 DUNLEAVY

To remain at the forefront of medical education, HMS must anticipate the physician of the future, practicing in an environment rich with cognitive support resources powered by artificial intelligence tools.

GEORGE Q. DALEY AB ’82, MD ’91, DEAN OF HARVARD MEDICAL SCHOOL AND CAROLINE SHIELDS WALKER PROFESSOR OF MEDICINE


One of the key initiatives backed by the fund is the Artificial Intelligence in Medicine (AIM) PhD track, which welcomed its inaugural class in 2024. Part of the Biomedical Informatics PhD program, it focuses on recruiting students right out of college who have strong training in computer science, engineering, mathematics, and other quantitative disciplines, as well as an interest in biology, medicine, and improving clinical care, explains Isaac Kohane, the Marion V. Nelson Professor of Biomedical Informatics, chair of the Department of Biomedical Informatics in the Blavatnik Institute at HMS, and co-director of the Ivan and Francesca Berkowitz Family Living Laboratory Collaboration at HMS and Clalit Research Institute.

Kohane notes that achieving the promise of AI in medicine requires a workforce skilled in developing safe, effective models for everyday clinical use. The PhD track can address this by training students in key AI health care applications, such as speeding up diagnoses, personalizing treatments, and predicting and preventing diseases to reduce costs.

“To remain at the forefront of medical education, HMS must anticipate the physician of the future, practicing in an environment rich with cognitive support resources powered by artificial intelligence tools,” says George Q. Daley AB ’82, MD ’91, dean of HMS and Caroline Shields Walker Professor of Medicine. “The time to invest in building a pipeline of AI experts for health care is now. Support from the Dunleavy Foundation is essential in helping HMS to lead the way.”

 { Q+A }

Aiming to Enhance Health Care

The Artificial Intelligence in Medicine (AIM) PhD track focuses on training exceptional computational students to harness large-scale biomedical data and advanced AI methods to develop new clinical research tools. Inaugural AIM PhD students Andrew Zhou and Thomas Buckley share their experiences in the program and where they see AI improving health care.

Image
Andrew Zhou
ANDREW ZHOU
Image
Thomas Buckley
THOMAS BUCKLEY

What inspired you to join the AIM PhD track?
AZ: AI is a revolutionary technology that’s positioned to make some of its biggest impacts in health care and biological science. It’s crucial that those who drive these innovations have an intricate understanding of medicine. The AIM PhD track is unique because of its faculty expertise in both medicine and computer science, and the opportunity to train in this environment was too good to pass up.

TB: When I was an undergraduate [at UMass Amherst], I was interested in machine learning, electrical engineering, and computer science. I wasn’t even thinking of medicine. But late into my freshman year, I was diagnosed with melanoma. Coincidentally, we were doing a project on melanoma detection in my machine learning class. While the experience was stressful, it empowered me to look for ways I might apply computer science to help people in my community. I found the Summer Institute in Biomedical Informatics at HMS—and that served as a natural pipeline to the PhD program.

Can you describe what you’re working on?
AZ: I’m exploring ways that cutting-edge AI methods can assist in drug discovery, particularly for antibiotics. Drug discovery traditionally requires a lot of chemical screening and testing, which are labor intensive and expensive processes that don’t always lead to the most effective therapies. I’m focused on how we can best model molecules in order to better predict their potential as antibiotics. With this approach, we might not only design 
antibiotics—and medicine in general—more cheaply and effectively, but we might also discover creative new therapies.

TB: I’m working with [Assistant Professor] Raj Manrai [AB ’08] on using AI to advance medical diagnoses. We’re working with chatbots, large language models, and programs like ChatGPT to rigorously evaluate whether they can perform medical diagnosis, testing them with the hardest medical cases to see how well they perform against physician-based science.

Why is it so important to merge AI and health care?
AZ: In health care, every small improvement could result in meaningful gains in patient health and experience. For instance, physicians and medical providers see a lot of patients, and they need to document these visits in great detail. One pilot program being rolled out at Mass General Brigham is AI that can listen to the doctor-patient visit and then prepare notes. This frees the doctor to conduct the patient visit face-to-face rather than focus on typing on the computer. Beyond administrative tasks, AI could help medical practitioners make better clinical decisions given information from the patient’s chart.

TB: In America, it’s getting harder to see physicians, and we’re one of the most well-resourced countries in the world. Under-resourced areas have even less access to health care. Imagine if I could have taken a picture of my melanoma and gotten expert medical advice quickly. These types of AI applications will be important in the future.

How have you benefited from the PhD track? 
AZ: It’s been the coolest classroom experience I’ve had this semester. The faculty all bring their different perspectives, and we’ve enjoyed long, insightful discussions. As students, we’ve never been afraid to ask questions; it feels like speaking with friends.

TB: I love it. I think it’s a great option for technical people who want to get into medicine. My mentor has been amazing, helping me define ambitious projects and enabling me to collaborate with doctors. I’ve benefited enormously.

Read more about groundbreaking work in AI across the University.