A team of iSchool researchers has been awarded a prestigious allocation through the National Artificial Intelligence Research Resource (NAIRR) Pilot Program to advance cutting-edge research in artificial intelligence for healthcare. The project, titled “Fine-Tuning Open-Source Large Language Models to Support Real-World Patient–Clinician Communication in Laboratory Test Results Interpretation” (NAIRR250293), was reviewed and approved for full allocation by the NAIRR program.

The NAIRR Pilot, led by the U.S. National Science Foundation (NSF) in partnership with the U.S. Department of Energy (DOE) and major public and private sector collaborators, provides researchers nationwide with access to advanced computing, AI models, platforms, and educational resources. The program aims to accelerate innovation in trustworthy, interpretable, and open AI systems across science, engineering, and societal applications.

Reviewers noted the project’s strong alignment with NAIRR’s mission to promote open-source AI development and the responsible use of advanced computational resources.  Purdue University’s Anvil AI has committed to supporting the FSU project with an allocation of 2,000 NVIDIA H100 GPU-hours, enabling large-scale fine-tuning and evaluation of state-of-the-art open-source language models.

“This project reflects our commitment to making advanced AI truly meaningful at the point of care,” said Dr. Zhe He, Professor in the School of Information and Director of the Institute for Successful Longevity. “By fine-tuning open-source language models specifically for laboratory test communication, we aim to empower patients and caregivers with clearer, more trustworthy explanations and more tailored questions for their clinicians. I’m excited to co-lead this work with Dr. Balu Bhasuran, whose expertise in natural language processing and clinical AI is central to translating these models into real-world, patient-centered tools.”

The project is co-led by Dr. He and Dr. Bhasuran, Research Faculty at the eHealth Lab within CCI. Together, the team will focus on improving how patients and caregivers understand laboratory test results through AI-powered, patient-centered communication tools.

The work builds on ongoing federally funded research to develop LabGenie, a web-based system designed to generate personalized explanations and clinically relevant questions that patients can use when speaking with their healthcare providers. By fine-tuning open-source large language models for real-world clinical communication, the project aims to enhance transparency, trust, and shared decision-making in healthcare—particularly for older adults and caregivers navigating complex medical information.

The team’s findings and open-source models, datasets, and methods will be shared with the broader research community, supporting NAIRR’s national goal of expanding access to high-quality AI