Conference on Artificial Intelligence in Medical Imaging

Agentic AI systems and foundation models are rapidly becoming central to the future of medical AI research, with proposed applications across imaging, reporting, clinical decision support, workflow orchestration, patient communication, and real-world deployment. Yet the field is still early, and many claims remain difficult to evaluate. Performance may depend on benchmark design, data leakage, prompt sensitivity, model access, evaluation strategy, clinical context, and the degree of human oversight. 

This interactive CAIMI research session will use a “Hip vs Hype” format to critically examine emerging directions in agentic AI and foundation models. Rather than focusing only on what is new or exciting, the discussion will ask what evidence is needed to determine whether these approaches represent meaningful scientific progress, translational opportunity, or premature enthusiasm. 

Using polls, participants will vote on key topics in agentic AI and foundation models, including autonomous clinical agents, multi-agent workflows, AI-assisted reporting, patient-facing systems, model evaluation, safety monitoring, and governance. Each poll will be followed by brief commentary and audience discussion focused on research gaps, methodological limitations, clinical relevance, and opportunities for collaborative study. 

The goal of the session is to create a dynamic, audience-driven conversation that helps researchers distinguish promising frontiers from under-validated claims and identify rigorous research questions for the next phase of medical AI.