Understanding Algorithmic Deference: Pathways through Psychological Safety and Leader–Member Exchange

Organizations rely on algorithmic systems that shape how people perceive situations at work, prompting concern about how employees adapt their decision-making habits and sense of psychological safety. We explored these changes by developing a novel construct, Algorithmic Deference, that describes the inclination to defer to algorithmic recommendations. We deployed a mixed design. Qualitative interviews helped define the construct and its items, and quantitative data from a South Asian sample helped validate the construct and the proposed moderated mediation hypothesis. Findings indicate that it decreases psychological safety, which in turn decreases innovation, but positive leader relationships moderate this impact. We advance the algorithmic leadership literature by revealing a previously obscured mechanism in human–algorithm relationships. We offer leaders insights into creating healthy digital workplaces, accounting for the interpersonal consequences of automation. The research contributes by building the algorithmic deference as a hitherto unexplored construct that captures the changing nature of work.