Perks or perils? The dual effect of artificial intelligence-augmented human resource management digitization on employee performance

Abstract

This study aims to examine why artificial intelligence-augmented human resource management (AI-HRM) digitization has mixed implications for employee performance. By drawing on person–environment fit theory and sensemaking theory, we propose a contingent dual-pathway model in which AI-HRM digitization affects employee performance through job-related cynicism and innovative job behavior, with task discretion shaping the pathways. We tested the hypotheses by using three-wave time-lagged multisource survey data from full-time employees and their direct supervisors in China (N = 543 employee–supervisor dyads). AI-HRM digitization has no significant direct relationship with employee performance when the mediating mechanisms are considered. Instead, the construct functions through two opposing indirect pathways, that is, a negative pathway through job-related cynicism and a positive pathway through innovative job behavior, with task discretion moderating both pathways. Specifically, high task discretion can weaken the cynicism-based pathway but strengthen the innovation-based pathway, whereas low task discretion can strengthen the cynicism-based pathway but weaken the innovation-based pathway or make it nonsignificant. Our study advances a contingent dual-pathway model of AI-HRM digitization by showing that the same digital HR transformation mechanism can elicit defensive and proactive employee responses simultaneously. By identifying task discretion as a key boundary condition, the study can explain when AI-HRM digitization will undermine or enhance employee performance and thus moves beyond simple net-effect explanations for AI-HRM outcomes.