AI Agent & Human Behavior

TIMA combines generative AI with established human mobility theory to model how individual behavioral choices scale into population-level movement patterns. TIMA reproduces realistic travel patterns, heterogeneous activity spaces, and behavioral responses to disruptions such as the COVID-19 pandemic.

Policy Agent uses demographically representative LLM agents to simulate how people change their behavior in response to evolving disease risk and public health policies. Evaluated across Boston, Denver, and San Antonio, the framework reproduces broad city-level mobility responses without fitting to observed mobility data, providing a new way to model behavioral adaptation within population and epidemic systems.

We evaluate whether large language models can reproduce individual vaccination decisions using demographic, attitudinal, and media-exposure data. The study reveals both the potential and limitations of LLM agents for modeling real-world health behavior.