What I'm Working On

Most recently: How to measure alignment of AI in clinical decision-making of AI models vs humans and humans (e.g. clinicians) vs. other humans (e.g. patients). I continue to work with the Coordinating Center of the Undiagnosed Disease Network where we are ramping up efficiency through the use of AI, including providing self-service solutions for patients. Also, the cases of several of the patients that we were not previously able to diagnose have yielded to new genomic insights, also powered by machine learning (e.g. splicing variants in compound heterozygous cases.). I am also quite busy plowing through submissions to NEJM AI and discussing the more interesting ones with the rest of the editors. Finally, I am revisiting a project I undertook in the 1990's, how to diagnose kids from their growth patterns. But now with LLMs and far more pediatric data.

Current Focus

NEJM AI Development

Building editorial processes and community around NEJM AI. Establishing new standards for publishing AI research in medicine with focus on rigorous evaluation and real-world clinical validation.

High priority

LLM Clinical Integration

Exploring safe and effective integration of large language models into clinical workflows. Focus on alignment techniques, and evaluation metrics for healthcare applications.

Experimental phase

Dataset Shift Research

Investigating AI model performance across populations and time and variations in human values/utilities.

Ongoing

AI Safety in Healthcare

Developing frameworks for monitoring and mitigating risks in clinical AI.

Research

Rare Disease AI

Applying AI to improve diagnosis and treatment of rare diseases.

Clinical trials

Looking for Collaboration

Researchers and clinicians I'd love to work with

Doctor volunteers

We are looking for clinicians who are willing to review brief clinical cases so that we can compare their decisions to those of AI models.

Comparing preferences between clincians and AI.

Regulatory and financial implications of clinical alignment

Bias mitigation.

Healthcare Data Scientists

Data scientists who understand healthcare data complexities and are passionate about advancing the healthcare at the frontlines.

EHR data analysis expertise

Primary prevention in pediatrics

Multi-institutional data integration

International Partners

Building global partnerships to understand AI in healthcare across different healthcare systems, regulatory environments, and cultural contexts.

Cross-cultural AI research

Comparative healthcare systems

Global health informatics

Get in Touch

About

Isaac "Zak" Kohane, MD, PhD is Professor and Chair of the Department of Biomedical Informatics at Harvard Medical School and Editor-in-Chief of NEJM AI. Over his 30+ year career, he has pioneered the application of artificial intelligence and computational methods to clinical medicine, from early work on temporal reasoning in medical expert systems to current research on large language models in healthcare. Trained as a pediatric endocrinologist with a doctorate in computer science, he has bridged the gap between computational innovation and clinical practice, shaping how healthcare settings can become living laboratories for translational research. His work spans patient data sovereignty, clinical decision support systems, and the ethical deployment of AI in medicine. He has authored over 500 publications and mentored dozens of researchers who have gone on to lead AI initiatives across academia and industry.

Commercial Involvement