Skip to content

Research

The problems we work on.

Most of our work circles a single question: what does it take for humans and artificial intelligence to genuinely operate as one team? We study the cognitive machinery of human-AI collaboration (shared mental models, trust, situational awareness, and information flow) in the high-stakes environments where getting it wrong costs lives, missions, and trust in the technology itself. We certainly don't have all the answers yet, but we think these are the right questions.

Core Thrust · 01

Human-AI Team Design

How team roles, composition, and interaction shift when a teammate is artificial.

The foundation of the lab's work. We study how AI agents move from tools to teammates: how team composition, communication, leadership, and interaction design change when one or more teammates is artificial. This work spans empirical lab studies, computational simulations, and field-informed frameworks, and includes a systematic review of the empirical human-autonomy teaming literature that colleagues across the field have generously built upon.

Core Thrust · 02

Team Cognition & Shared Mental Models

The moment a human mind and an AI agent lock into shared understanding.

High-performing teams think together. We measure and model shared mental models, team situational awareness, and transactive memory in teams that include AI, asking what an AI teammate must know, share, and anticipate for the team to develop genuine shared cognition.

Core Thrust · 03

Trust in Human-AI Teams

How trust forms, breaks, spreads, and recovers across human-AI teams.

Trust is the load-bearing wall of human-AI collaboration. We study how trust develops and is violated in high-risk tasks, how trust and distrust spread contagiously through multi-human multi-AI team constellations, and how context and team role reshape the consequences of AI failure. Funded by the Army Research Office to study compromised AI teammates.

Core Thrust · 04

AI Ethics & Responsible AI

What ethical AI teammates owe their human counterparts.

AI teammates make decisions with moral weight. We empirically investigate how the ethicality of AI actions shapes trust and performance, how AI should explain and justify ethically-charged decisions, and how to design adaptive autonomy that keeps AI behavior aligned with team ethics, work that reviewers and colleagues have been kind enough to recognize with several best paper awards.

Core Thrust · 05

Intelligent Information Sharing

The right information, to the right teammate, at the right moment.

An AI teammate that says everything is as useless as one that says nothing. We model what information humans actually need from AI teammates, when sharing builds team cognition versus eroding it, and how information-sharing strategy drives trust: the difference between an AI that interrupts and an AI that anticipates.

Applied Domain · 06

Defense & Security

Human-machine teaming where the stakes are highest.

Command-and-control, battlefield AI, and compromised-teammate scenarios. The lab's ARO-funded work develops methods to prevent, identify, and mitigate compromised AI teammates using shared situation awareness and mental model accuracy. Invited briefings include West Point's Technical Advisory Panel on AI in warfare and the National Academies Board on Human-Systems Integration.

Applied Domain · 07

Healthcare

Ethical adaptive AI for clinical teams.

Healthcare teams are adopting AI faster than we understand its team-level effects. We study adaptive autonomy and ethical AI teammate design for clinical environments, where an AI teammate's judgment intersects with patient safety.