Algorithmic Green Human Resource Management and Employee Green Voice: The Mediating Roles of Perceived AI Fairness and Surveillance Anxiety with Green Psychological Safety as a Moderator

Abstract

Algorithmic Green Human Resource Management (Algorithmic Green HRM) is reshaping how organizations track, evaluate and incentivize employee pro-environmental behaviour. Yet the same algorithmic infrastructure that fairly distributes green opportunities can also generate surveillance anxiety, turning employees away from green voice and into defensive green silence. Utilizing theories of Organizational Justice, Psychological Safety and Conservation of Resources, the current study proposes and tests a dual-pathway, dual-outcome moderated-mediation model, in which Algorithmic Green HRM affects Green Voice Behavior (GVB) and Defensive Green Silence (DGS) via a bright pathway of Perceived AI Fairness (PAIF) and dark pathway of Surveillance Anxiety (SA). Moreover, Green Psychological Safety (GPS) moderates both first-stage paths. A multi-source, three-wave study was undertaken with 438 employees and their supervisors in five industries of Pakistan. Data were analyzed with partial-least-squares structural equation modelling (PLS-SEM), covariance-based confirmatory factor analysis, and bias-corrected bootstrap mediation. Results show that Algorithmic Green HRM produces both fairness perceptions (β = 0.547) and surveillance anxiety (β = 0.418); fairness positively predicts voice (β = 0.474) and negatively predicts silence (β = -0.214), whereas anxiety positively predicts silence (β = 0.402) and negatively predicts voice (β = -0.182). Green Psychological Safety amplifies the bright fairness-to-voice path and buffers the dark anxiety-to-silence path, generating significant indices of moderated mediation (0.062 bright; -0.053 dark). The model explains 46.2% of variance in voice and 34.1% in silence, outperforming nine benchmark studies published between 2020 and 2026. The paper contributes a theoretically integrated dual-pathway account of algorithmic management in sustainability HRM, validates measures of green-specific voice, silence and surveillance anxiety, and offers a five-tier policy framework for ethical algorithmic green HRM in emerging-economy contexts.