Attitudes toward artificial intelligence in the workplace: a review of antecedents, consequences and mechanisms

Counts differ because each database indexes a different set of publications. We treat OpenAlex as the canonical count; Google Scholar is not shown (no API, and crawling it violates its ToS).

Purpose This paper aims to synthesize existing empirical research on employees’ attitudes toward artificial intelligence (AI) in the workplace. It organizes existing evidence around five attitude dimensions – perceived usefulness, aversion, trust, perceived threat and ethical perceptions – to clarify the current state of knowledge on employee responses to AI integration and to develop theoretical insights for future research and practical implications for management practices. Design/methodology/approach The paper analyzes 43 high-quality empirical studies retrieved from the Web of Science and China National Knowledge Infrastructure Chinese Social Sciences Citation Index databases. This review integrates six theoretical perspectives to categorize and synthesize the antecedents, outcomes and mechanisms of employee AI attitudes. Findings The reviewed evidence suggests that positive attitudes are generally associated with adaptive behaviors, while negative attitudes often relate to resistance, strain or withdrawal. In addition, this study identifies and seeks to reconcile several areas of inconsistent findings, including the dual effects of AI threat, divergent fairness intuitions and the double-edged role of anthropomorphism, proposing possible boundary conditions such as task attributes and interaction modes. Originality/value This paper develops a systematic “antecedents–attitudes–outcomes” framework, contributing to the limited but growing body of reviews on workplace AI attitudes. Beyond synthesizing existing conclusions, this study identifies potential conflicts within the literature and attempts theoretical reconciliation by introducing moderating variables. Furthermore, it offers a critical analysis of complex attitudinal states and identifies key future research directions, including generative AI, causal inference designs and the risks of over-trust.