A five-year, $21.5 million National Science Foundation grant will power a major collaboration between Penn, the University of Delaware, and more than 40 other academic, industry, and government partners. This new partnership aims to launch the AI Institute for Human-AI Cooperation (HAIC). The institute is focused on integrating artificial intelligence into health care in ways that support, rather than replace, human decision-making.
Cross-sector collaboration to advance AI in health care
The HAIC institute will unite over 40 stakeholders to develop AI systems designed to aid human clinicians. These systems are not intended to take their place. Eric Eaton, who serves as HAIC's managing director, is a professor in Penn's computer science department. He stressed the importance of creating AI tools that enhance clinical judgment. These tools should not eliminate human input. His vision is to construct a national research hub. In this hub, AI and humans can coexist and collaborate in meaningful ways.
Cathy Wu, HAIC's director, explained the institute's goal. The goal is to turn artificial intelligence into a flexible and responsive teammate. These systems will offer dynamic support that adapts to specific patient needs. This adaptation will open new possibilities in health care. It will also train the next generation of AI leaders. This shift, Wu said, could redefine the roles and responsibilities of AI in clinical environments.
Innovative data and technology approaches for personalized care
HAIC will use advanced tools to analyze complex, multiscale, and multimodal data in ways that make rehabilitation more effective and tailored. The institute plans to employ AI-powered digital twins, which are virtual replicas of patients, to simulate various treatment scenarios and help clinicians optimize their approaches before real-world implementation.
Penn’s GRASP Lab will be a key player in the HAIC initiative, contributing its strengths in foundational AI, machine learning, and robotics. Eaton outlined several research areas the lab will focus on, including multimodal reasoning of time-varying processes, adaptive and continual learning systems, and methods to improve human-AI interaction. These areas are central to the institute's work and will help bridge the gap between AI innovation and clinical application.
Enhancing the patient experience through AI
Austin Brockmeier, HAIC's deputy director, emphasized the role of AI as a supportive tool in the patient journey. He explained that AI could offer a second opinion, analyzing data in ways humans cannot, thereby increasing the depth and breadth of information used in treatment planning. Brockmeier also highlighted the potential for AI to improve recovery outside the clinic, through home-based monitoring and adaptive guidance during exercises, which can significantly enhance the patient experience.
The institute plans to form a triad involving AI, patients, and clinicians, creating a more collaborative model of care. Wu praised the diverse expertise contributed by collaborating institutions, noting that HAIC's foundational AI work could lead to significant economic and societal benefits beyond the health care sector. She described the project as a collective effort that combines varied strengths to drive research and real-world impact.