Leader-Follower Hierarchical Role Effects on Individual and Shared Trust in Mixed Human-AI Teams

Human-AI Teams (HATs) are evolving into increasingly complex multi-human multi-agent structures, including those with hierarchical leadership and role accountability. Conventionally, humans have primarily taken leadership positions, directing and supervising autonomous agents. However, emerging HAT contexts now enable humans to assume a wider range of hierarchical roles within teams, including both leader and follower positions. Despite these developments, empirical research on non-dyadic HATs remains limited, leaving important questions regarding trust and teamwork in complex team structures largely unexplored. In this study, we investigate how hierarchical roles differentially shape trust in AI teammates and trust in the team as a whole, as well as how role-based trust divergence associates with overall team outcomes. We find that trust varies significantly by the trustor’s role, yet shared trust and reduced trust divergence across teammates are critical for better team performance. Accordingly, it is important to support shared team cognition and mitigate trust divergence.