STRONG

The Signatures of Success in Human-Agent Teams

Funded by the Department of the Army

Background

 U.S. Army squads are facing two transformative changes. Each with the potential to significantly augment military capability. First, advances in AI and computer science are enabling intelligent, autonomous agents to join military teams. This invites new theorizing on how the processes and outcomes of human autonomy teams extend or amend current theories on human teams.  Second, wearable sensors and advanced real time processing are making it possible to leverage human physiological and behavioral data streams to understand and enhance performance not only in human teams but to also feed data into autonomy systems to enable smooth human-autonomy collaboration. Clearly, enabling human-autonomy teaming requires foundational knowledge spanning time scales and units of analysis, from brain signals to inter-team cognition.This project focuses on the first of these two  challenges – extending extant theories of human teams to human-autonomy teaming. 

This project has three research objectives (ROs):                      

    • RO1: Identify the team processes and properties that predict success in human-autonomy teams.
    • RO2: Isolate the unique effects of autonomy (versus humans) on team processes and outcomes. 
    • RO3: Identify learning signatures associated with successful human-autonomy teaming.                                                        

These objectives advance theoretical understanding on human-autonomy teaming along two dimensions.  First, we move current thinking from dyadic human-autonomy interactions to complex, systemic human autonomy teaming. Second, we move from the extant – largely static – human teaming theories to dynamical network theories of the impact of human autonomy teaming on team processes and outcomes.