How safe is too safe? A dual-pathway framework for understanding the dark side of psychological safety

Psychological safety has become one of the most celebrated constructs in organizational behavior, consistently linked to voice, learning, and team effectiveness. This paper challenges the assumption that reducing interpersonal risk is universally beneficial. Psychological safety operates through two competing pathways: one that enables expression and participation, and one that weakens the interpersonal evaluative pressure, self-monitoring, and norm adherence that sustain coordination and performance. This second mechanism is termed the regulatory loss pathway. The pathway operates through three behavioral channels: the loss of functional filtering, effort withdrawal, and norm erosion. Drawing on self-regulation theory, accountability research, and the social loafing literature, the paper develops a dual-pathway model that specifies when psychological safety helps and when it hurts. Evidence from military operations, aviation, and high-reliability organizations shows that some of the constraints psychological safety removes serve as structural conditions for coordinated and timely decisions. The model predicts that the relationship between psychological safety and performance is nonlinear, with moderate levels producing the best outcomes and very high levels generating process losses, particularly where coordination demands are high and external accountability is weak. The framework also specifies how hierarchy and power asymmetries shape both pathways. Seven testable propositions and key boundary conditions are developed, with implications for high-stakes organizational design.