The Harvard Data Science Initiative (HDSI) in conjunction with Harvard University Information Technology (HUIT) and the Office for the Vice Provost for Research (OVPR) has selected two projects for support through the Innovation API Fund, providing OpenAI API credits to accelerate research efforts across Harvard University. Learn more about the fund and how to apply on the Innovation API Fund page.
Validating Large Language Model (LLM) Vignettes for Provider Competency Assessment
At the Harvard T. H. Chan School of Public Health, Kevin Croke is investigating how large language models can be used to assess provider competency at scale. This project evaluates the reliability of LLMs in extracting key diagnostic features from simulated provider–patient interactions, comparing automated and human-coded assessments to create a faster, more scalable, and globally applicable evaluation framework.
Characterizing the Modularity by an AI + Human Approach
Led by Subhabrata Sen at the Department of Statistics, this project explores how frontier LLMs like GPT-5 can contribute to original mathematical research. Focusing on the problem of modularity in random graphs, it tests whether AI can help close existing theoretical gaps through collaborative reasoning—advancing both mathematical knowledge and our understanding of AI’s potential in fundamental research.