When humans stop thinking: tackling the silent threat of AI complacency in service operations

Purpose – This research investigates the phenomenon of AI complacency – The employee's tendency to intentionally neglect validating AI-generated output even in the presence of systematic errors. It identifies the lack of monitoring accountability as the underlying driver of this phenomenon, assesses its consequences, and offers strategies for mitigation. Design/methodology/approach – Grounded in contingency theory, this research employs six experimental studies (N = 1, 370 participants), including one study with service employees across industries (N = 160 participants), to examine how insufficient monitoring accountability facilitates the emergence of AI complacency. The research explores both the causal mechanisms and the boundary conditions that modulate this effect. Findings – The results show that the primary driver of AI complacency is the lack of accountability in monitoring AI-generated outputs. This complacency leads to detrimental work-related outcomes, such as increased commission errors and a diminished willingness to evaluate AI-generated outputs critically. The research also identifies situational factors that exacerbate and buffer these effects. Practical implications – The findings highlight the critical need for organizations to implement systemic accountability frameworks that ensure employees actively engage with and oversee AI-generated output. Originality/value – This research is among the first to examine AI complacency in the context of service provision empirically. It provides a theoretical framework, robust empirical evidence, and practical recommendations for improving Employee-AI collaboration in service provision, contributing to both academic discourse and managerial practice.