Foundation models (e.g., large language models and multimodal generative systems) offer transformative potential for biomedical and clinical tasks. Yet, their application in public health — which involves population-level decision-making, surveillance, and equity-sensitive interventions — remains underdeveloped. Despite many methodological studies, there is limited translation into practical use, low generalizability across populations, and poor interoperability with existing health data systems. This workshop will convene researchers and practitioners to critically assess the current landscape, identify barriers to real-world impact, and chart a research agenda that connects foundation model capabilities with public health needs. The half-day workshop will include invited talks, contributed short presentations, and structured discussions, concluding with a community-driven roadmap for evaluation frameworks, hybrid modeling (including agent-based integration with generative AI), and interoperable deployment strategies. We aim to foster collaboration between AI methodologists, public health informaticians, and policy makers, strengthening the field’s scientific foundations and translational impact. Ultimately, the workshop aims to catalyze collaborative initiatives (e.g. shared benchmarks and consensus evaluation guidelines) to be disseminated through the AIME community, fostering scientifically rigorous and operationally meaningful adoption of foundation models in public health.

This workshop invites contributions from experts and researchers across diverse disciplines, such as computer science, epidemiology, environmental science, medical informatics and public health. 

We welcome submissions on, but not limited to:

• Foundation models for public health surveillance and/or outbreak detection

• Multimodal generative models for epidemiological and environmental data

• Generalization & fairness across diverse population groups

• Interoperability with public health information systems and standards

• Evaluation metrics and validation frameworks tailored to public health impact

• Ethical, governance, and equity considerations in population-level AI

• Integration of agent-based models with generative AI for policy simulation

• Case studies and practical deployments in public health agencies

Best papers will be considered for publication in the Machine Learning and Knowledge Extraction  journal with the possibility of APC waivers.