Who Manages the AI Manager? HR Governance, Employee Trust, and Workplace Outcomes in Autonomous Work Systems

Abstract

Purpose: This study investigates how AI managerial autonomy affects employee trust and organizational justice, whether HR governance quality moderates these relationships, and whether trust mediates the effects of governance on employee outcomes (organizational commitment, job satisfaction, and turnover intention) in autonomous work systems. Design/Methodology/Approach: A time-lagged survey was conducted with 387 employees from 16 organizations across Karachi, Lahore, and Islamabad (banking, telecom, manufacturing, IT). Partial least squares structural equation modelling (PLS-SEM) was used with bootstrapped confidence intervals. A priori power analysis (G*Power) confirmed adequate sample size. Findings: AI managerial autonomy is negatively associated with employee trust (β = –0.31, p < .001) but shows no significant association with organizational justice (β = –0.09). HR governance quality moderates the AMA–trust relationship (β = 0.29, p < .001) but not the AMA–justice relationship (β = 0.07) Algorithmic transparency is the strongest governance predictor of trust (β = 0.44). Trust mediates governance effects on commitment (indirect β = 0.31), satisfaction (0.26), and turnover (–0.22). Sector heterogeneity is observed: manufacturing employees are most negatively affected. Originality: This study is one of the first to empirically examine who governs the AI manager. It advances the Human-AI Governance Framework (HAGF) by positioning HR governance as a moderating institutional mechanism, distinct from prior design focused frameworks. It also provides realistic non-significant justice findings, highlighting that trust, not justice, is the primary casualty of ungoverned algorithmic management.