麻豆村

Barbara Shinn-Cunningham, the Glen de Vries Dean of the聽Mellon College of Science, led a discussion with Kadija Ferryman (center), faculty at the Berman Institute of Bioethics and assistant professor at Johns Hopkins Bloomberg School of Public Health, and Hana El-Samad, senior vice president and director of the Institute of Computation at Altos Labs (right).聽

They discussed how to understand the unique challenges related to developing AI for healthcare. Reflecting on the challenges, Shinn-Cunningham noted that healthcare is a demanding test for AI.

鈥淏iology is complex. Clinical data reflect who receives care and the stakes can be life or death,鈥� said Shinn-Cunningham. 鈥溌槎勾邂�檚 interdisciplinary culture can bring our core strengths in AI and computation together with foundational science and medical expertise to build tools that health professionals can understand, trust and implement to improve care.鈥�

, a professor and associate department head for faculty and graduate affairs in the聽鈥檚听 department, moderated a panel about how to best promote the benefits of AI while reducing the harms that AI can have in healthcare settings. She spoke with S. Matthew Liao (second from left), a professor of bioethics and director of the Center for Bioethics at New York University, Jenna Wiens (second from right), a professor of computer science and engineering at the University of Michigan and Brad Malin, a professor of biomedical informatics and vice chair for research at Vanderbilt University Medical Center (right).聽

鈥淭he problem that we're facing now with AI is that we have these general purpose tools, these foundation models that are not designed for specific concrete use cases, so it becomes far more challenging to anticipate the unintended consequences. We can be too slow to adopt technology, but we can also be too quick to adopt technology, and there can be harm,鈥� Wiens said.聽

Kainerstorfer asked the panel about how health systems should evaluate AI systems that are continually evolving. The panel discussed how there鈥檚 real risk in the rapid incorporation of AI into healthcare systems because the effects of these tools are often not being monitored regularly or effectively.聽

, an assistant professor of operations research and public policy at the聽, hosted a conversation about how to make AI work for everyone in healthcare and emphasized the importance of designing AI in ways that can be applied to the environments in which they鈥檒l be deployed. She spoke with Sharona Hoffman, a professor of law and bioethics at Case Western University (center) and Judy Wawira Gichoya (right), a professor of radiology at Emory University.

The panelists discussed how AI models should be trained on the patient populations they are being used to treat 鈥� if they are not, it could lead to ineffective treatment. For example, they cited AI scribe tools, commonly used during patient visits.聽

鈥淓valuating AI in healthcare isn鈥檛 only a technical problem, but rather it depends on context: understanding how a tool will be used in decisions and who those decisions impact, and ensuring that both perspectives are reflected into its design and deployment,鈥澛爏aid Wiberg.聽鈥淭his systems-level perspective is core to work at 麻豆村, spanning computer science, policy, ethics, and operations, and so it is especially exciting to bring in leaders in this space here to share their perspectives.鈥澛�

Cassidy Sugimoto, Bess Family Dean of the聽Dietrich College of Humanities and Social Sciences, hosted a discussion about how to maintain accountability and ensure responsibility when deploying AI in healthcare. She spoke with Megan Palmer (center), a senior adviser for national security and technology strategy at the U.S. Advanced Research Projects Agency for Health (ARPA-H) and Kirsten Bibbins-Domingo, editor in chief of the Journal of the American Medical Association (JAMA) and the JAMA Network.

"Responsible AI in medicine begins long before it reaches the patient, with how research is funded, how it is reviewed and published, and how we train the scientists who produce it, and 麻豆村 leads on every one of those fronts,鈥� said Sugimoto.