Human–AI teamwork.

Knowing who knows what is only useful if a team can bring that knowledge into a decision. We build AI agents that maintain knowledge networks across teams. I study what information an AI should share, when people need it, and how they use it to form a collective judgement.

In my studies, people changed their behavior to complement an AI teammate. I study how that influence develops and what it means for the relationships among human teammates. That work informs how I design AI’s role in a team, especially when people and AI disagree.

An AI’s role may need to change as the situation changes. We build adaptive systems for settings including cyber incident response and healthcare. I study what an AI needs to consider before changing its behavior, so people can still understand the situation and coordinate a response.

I want AI to catch and block cognitive risks before they damage teamwork. We build agents for offensive and defensive cognitive warfare research, studying how failures and adversarial influence can disrupt a team. I use that work to design AI support and training that protect how people think and work together.