Artificial Agents as Leaders and Followers: An Integrative Role-Congruity Review

Artificial agents such as algorithms, virtual agents, and robots increasingly occupy organizational roles as leaders, followers, and partners. However, evidence on how people evaluate these nonhuman role occupants is mixed. We systematically review 107 empirical studies across management, psychology, engineering, and computer science through a role-congruity lens. A consistent pattern emerges: Artificial agents in leader roles face role incongruity, driven largely by perceived deficits in morality (e.g., fairness, trustworthiness). Agents in follower roles are evaluated as largely congruent with role expectations. Agents in partner and teammate roles show conditional congruity that depends on task context and design-based social cues (e.g., appearance, voice). We integrate these patterns into the Artificial Agent Role Congruity (AARC) model, a conceptual model that extends role congruity theory to artificial agents: perceived role (in)congruity occurs when observers compare role prescriptions with the attributes they ascribe to an agent from its design-based cues along a task dimension (competence, assertiveness) and a social-emotional dimension (warmth, morality). Based on this synthesis, we outline theoretical extensions, an agenda for future research, and guidance for managers and designers deploying artificial agents in organizational roles.