To the YSM Community:
As noted in the January Beyond Sterling Hall on Communication 3.0, over the coming months I will highlight efforts to realize cross-cutting themes in our Yale School of Medicine (YSM) Strategic Plan. Investment in the development and implementation of artificial intelligence (AI) permeates plan goals throughout education, clinical, and research missions.
AI in Education
Under Education Strategic Goal 2, we aspire to “Develop educational opportunities in AI and technology.” To this end, YSM Deputy Dean for Education Jessica Illuzzi, MD, established the Educational Technology & Innovation (ETI) team in late 2024 to provide strategic leadership, coordination, and oversight for the educational use of emerging technologies such as AI in medicine.
ETI is helping faculty, students, and staff learn to optimize the use of emerging technology to improve teaching and learning. This includes integrating AI topics into the medical school curriculum, tackling ethical considerations, selecting language models for clinical reasoning, and using ambient scribes for clinical documentation. Educational sessions address practical skills such as prompt engineering, using custom chatbots, and applying AI in assessment and feedback. The Technology & Innovation in Medical Education (TIME) series provides workshops, lectures, and panel discussions on emerging technology.
ETI is also developing AI-enabled tools to improve quality, specificity, and consistency of formative feedback and competency-based assessment. For example, as students practice clinical skills, an ambient scribe converts verbal feedback from faculty into written narratives that students can use to guide future learning. Students can also practice communication skills using an AI-powered patient avatar, and a conversational chatbot trained on the YSM curriculum enables students, staff, and faculty to identify content anywhere in the curriculum.
ETI is collaborating across the school and health system to help YSM expand education on the use of AI in clinical settings to assist and improve patient care. As part of these broader efforts, medical students may take an advanced elective in the department of Biomedical Informatics and Data Science (BIDS) on applying programming, data science, and informatics to diverse clinical contexts. YSM has developed the first curriculum on the use of ambient scribes for clinical documentation for medical students as well as the first curriculum for residency and fellowship programs in the country.
Many graduate programs are incorporating AI into their curriculum. Recently, BIDS received approval to establish a new online Master of Health Science (MHS) Program in Medical AI. This new program targets “students with a technical background (computer/data science, engineering, statistics) who will be interested in both expanding their technical AI skills, and acquiring new domain-specific expertise in the application of AI in healthcare, including application areas, regulatory and other legal issues, and an understanding of the operation of healthcare systems.” It expands on the 16-week certificate program in medical software and biomedical informatics.
Clinical AI
Clinical Strategic Goal 5 calls for YSM to “Advance the application of precision medicine / (AI) in the academic health system, leveraging research to improve diagnosis and treatment.” Advances in AI offer the opportunity to accelerate diagnosis and tailor therapies to patients. YSM faculty such as Sanjay Aneja, MD, and Melissa Davis, MD, MBA, have been applying AI to improve prognostic prediction in patients with cancer and accelerate diagnoses of life-threatening thrombotic events, respectively. Similarly, in digital pathology, the application of AI and machine learning promises to lead to new diagnostic and prognostic biomarkers.
Implementation of AI and other forms of personalized medicine such as genetic/genomic risk prediction require that we iterate on “use cases to develop processes and infrastructure for personalized medicine across the health system, including approaches to patient identification, clinician and patient education, data sharing, clinical decision-making support, implementation, and outcome tracking.” Here Rohan Khera, MD, MS, and Yong-Hui Jiang, MD, PhD, and teams are leading interdisciplinary teams to pilot the implementation of AI to identify structural heart disease from electrocardiograms (ECGs) and the application of genetic screening for familial hypercholesterolemia (FH) in the clinic, respectively.
Responsible development and deployment of AI in clinical care requires disciplined oversight. Under the leadership of Lee Schwamm, MD, associate dean for digital strategy and transformation and senior vice president and chief digital health officer, YSM and Yale New Haven Health System, we have developed a healthcare AI governance process. Over the last six months, workgroups have reviewed 32 proposals in key areas such as clinical workflow, privacy and security, ethics and equity, and operational considerations. A Digital / AI Implementation Advisory (DAIAD) Committee performs a comprehensive review of the assessments and provides lifecycle oversight of AI solutions, ensuring rigorous vetting and risk evaluation prior to deployment. Pending the development of a website with general FAQs and instructions, developers of clinical AI can access intake forms at this link.
Research Using AI
Our strategic plan calls for YSM to “Lead in the development of data science, AI, and bioinformatics methods and applications in biomedical and clinical research” in Research Goal 4.
On March 26, the Yale Clinical and Translational Science Award (CTSA)-funded Yale Center for Clinical Investigation (YCCI) and the department of BIDS held a full-day Medical AI Symposium with informatics leaders from various east coast CTSA hubs (e.g., Cornell, University of Florida, Harvard, University of Massachusetts, Northwestern, University of Pennsylvania). In-person attendees numbered 240 and another 100 from various states attended online. The group shared experiences with new research applications, utilization of AI models across institutions, and the development of foundation models for health sciences. At Yale the development of foundation models based on electronic health records (EHRs) has been enabled by the creation of a secure GPU environment in the Hopper cluster, as well as participation of Yale faculty in the development of Epic’s largest EHR-based foundation model for medicine.
Our basic science departments are also rapidly growing efforts in AI through cross-university collaboration as well as strategic recruitment. While it is impossible to name the many faculty engaged in AI and computational biology, I will highlight a few recent recruits here. Carl-Mikael Suomivuori, PhD, in Pharmacology is using computational tools and machine learning structure prediction to design more effective drugs. Anupama Jha, PhD, (Genetics) is using predictive machine learning to study 3D genome architecture, downstream gene regulation, and how it is impacted in disease states. Daniel Levenstein, PhD, (Neuroscience) is developing AI systems that mimic spontaneous activity in the brain during sleep and offline learning to better understand how it supports learning. Armita Nourmohammad, PhD, (Immunobiology) is integrating approaches in machine learning, information theory, statistical physics, and control therapy to study adaptive immune responses.
Long-standing faculty member Steve Kleinstein, PhD, and his team have developed high throughput measurements of adaptive immune receptor repertoire sequencing (AIRR-seq) experiments to understand and predict individual differences in immune status and vaccine response. John Tsang, PhD, MMath, director, Yale Center for Systems and Engineering Immunology (CSEI); investigator and Yale lead, Chan Zuckerberg (CZ) Biohub NY, is applying machine learning, quantitative modeling, and experimental methods, including high-dimensional, longitudinal immune monitoring of human cohorts throughout the lifespan and around the globe, to understand individual variation in immune responses. Rong Fan, PhD, recently appointed director of the Yale Center for Human Spatial Biomedicine, has received funding through the CZI Biohub NY for his proposal, “decoding and engineering immunity through spatial multi-omics.”
Leveraging AI to Communicate
So how does one keep up with all that is going on? Here too, AI plays a role. Our communications team is using AI to drive the architecture of web content to optimize the ability to surface trusted information. They are developing an AI-powered chatbot to provide faculty, staff, and visitors with a conversational way to find information across our websites. AI automates the tagging and categorization of research projects using standardized MeSH vocabularies and connects them to related people, publications, and opportunities. The communications team has piloted the use of AI to generate summaries of articles. The team is now developing a tool “that lets any tagged Yale author review, approve, and publish summaries of their own work; auto-generates summaries for any paper with a YSM first or last author; sends the senior author a review notification; enforces an attestation checkbox; and adds an on-page AI disclosure.” The goal is to roll this out by the end of May 2026.
As we advance these efforts, we will prioritize responsible governance, ethical rigor, and measurable impact across our missions. Thank you for your creativity and vision as we build the infrastructure, skills, and culture needed to realize the promise of AI for our community and those we serve.
Sincerely,
Nancy J. Brown, MD
Jean and David W. Wallace Dean of Medicine
C.N.H. Long Professor of Internal Medicine