Strategic Multi-Agent Interaction with Human Language

As AI systems increasingly communicate, negotiate, and make decisions on behalf of humans, understanding their behavior in strategic environments becomes increasingly important. In this talk, I will present a line of work on predicting how humans and AI agents behave in language-based strategic interactions, including behavioral simulation and synthetic data, as well as agent evaluation and development. I will introduce GLEE, a framework for studying these interactions, and explore how the availability of different AI agents can shape individual decisions, strategic outcomes, and broader ecosystems of interacting humans and agents.