Autonomous AI managers erode worker trust, but good HR governance — notably algorithmic transparency — offsets the damage and raises commitment and satisfaction while cutting turnover intent; manufacturing employees suffer the most.

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

Dr. Farah Naz, Attique Ur Reman, Dr. Sara Sohaib, Ruby Usman, Muhammad Zohaib Saleem · February 28, 2026 · ACADEMIA International Journal for Social Sciences

openalex correlational low evidence 7/10 relevance Summary only summary available; pdf_status=paywall DOI Source PDF

Structured author observations

Linked only from stored provider relations; the raw author line above is never matched by name.

OpenAlex

Latest observation: August 25, 2026

Dr. Farah Naz provider ID
Attique Ur Reman provider ID
Dr. Sara Sohaib provider ID
Ruby Usman provider ID
Muhammad Zohaib Saleem provider ID

Semantic Scholar

Latest observation: August 25, 2026

Dr. Farah Latif Naz provider ID
Attique Ur Reman provider ID
Dr. Sara Sohaib provider ID
Ruby Usman provider ID
Muhammad Saleem provider ID

Greater AI managerial autonomy is associated with lower employee trust, but strong HR governance—especially algorithmic transparency—buffers that effect and, via trust, increases commitment and satisfaction while reducing turnover intentions, with manufacturing workers most adversely affected.

Citation observations

Cumulative provider counts captured on specific dates; providers are never combined.

0 cumulative citations

OpenAlex · Observed August 25, 2026

View corpus context

0 cumulative citations

Semantic Scholar · Observed August 25, 2026

View corpus context

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.

Summary

Main Finding

AI managerial autonomy (AMA) reduces employee trust, but HR governance quality can buffer that effect. Trust — rather than perceptions of organizational justice — is the key pathway linking HR governance to employee outcomes (organizational commitment, job satisfaction, turnover intention). Algorithmic transparency is the single strongest governance predictor of trust. Effects vary by sector, with manufacturing workers most negatively affected.

Key Points

Core relationships
- AMA → Trust: negative and significant (β = –0.31, p < .001).
- AMA → Organizational justice: non‑significant (β = –0.09).
Moderation by HR governance quality
- HR governance moderates the AMA→Trust relationship (interaction β = 0.29, p < .001): better HR governance attenuates the negative trust impact of AMA.
- No significant moderation for AMA→Justice (interaction β = 0.07).
Governance components
- Algorithmic transparency is the strongest single HR governance predictor of trust (β = 0.44).
Mediation by trust
- Trust mediates the effect of HR governance on employee outcomes:
  - Commitment: indirect β = 0.31
  - Job satisfaction: indirect β = 0.26
  - Turnover intention: indirect β = –0.22
Sector heterogeneity
- Manufacturing employees experienced the largest negative trust impact from AMA relative to banking, telecom, IT.
Contribution
- Advances the Human‑AI Governance Framework (HAGF) by positioning HR governance as an institutional moderator (who governs the AI manager), distinct from prior design‑centric approaches.
- Reports a realistic null finding for justice, indicating trust is more immediately affected by ungoverned algorithmic management.

Data & Methods

Sample
- N = 387 employees from 16 organizations across Karachi, Lahore, and Islamabad.
- Sectors: banking, telecom, manufacturing, IT.
Design
- Time‑lagged survey design.
Analysis
- Partial least squares structural equation modelling (PLS‑SEM) with bootstrapped confidence intervals.
- A priori power analysis (G*Power) confirmed adequate sample size.
Key statistics reported
- Direct and interaction betas and p‑values (see Key Points).
- Indirect (mediated) effect betas for trust on outcomes.

Implications for AI Economics

Institutional governance matters as much as algorithm design
- For firms and regulators concerned with AI in workplaces, investment in HR governance (especially algorithmic transparency) can reduce the negative trust consequences of granting autonomy to AI managers.
Worker outcomes and labor market dynamics
- Lower trust driven by AMA can reduce commitment and satisfaction and increase turnover intention, with potential productivity and retention costs — particularly acute in manufacturing.
Policy and organizational strategy
- Policies that require or incentivize algorithmic transparency and formal HR governance mechanisms around AI management may preserve worker trust and mitigate turnover risks.
Modeling AI‑labor interactions
- Economic models of AI adoption should incorporate institutional governance variables (HR rules, transparency, oversight) as moderators of worker responses, not only technical performance gains.
Heterogeneous effects
- Sectoral differences imply uneven labor market impacts of algorithmic management; transition costs and retraining policies may need sector‑specific targeting.
Research direction
- Future economic research should quantify the productivity vs. trust tradeoffs of AMA and evaluate cost‑effective governance interventions (e.g., transparency standards, participatory design) in randomized or quasi‑experimental settings.

Assessment

Paper Typecorrelational Evidence Strengthlow — Findings are based on self-reported, observational survey data without exogenous variation or experimental manipulation; time-lagging and statistical controls/bootstrapping help reduce bias but cannot rule out reverse causation, omitted confounders, or common-method variance, limiting causal inference. Methods Rigormedium — The study uses appropriate survey methods (time-lagged design), reports a priori power analysis, and applies PLS-SEM with bootstrapped inference and formal moderation/mediation tests, but relies on self-report measures, a non-random organizational sample, and PLS-SEM (which has known limitations versus covariance-based SEM), leaving potential measurement, selection, and confounding concerns. Sample387 employees from 16 organizations across three Pakistani cities (Karachi, Lahore, Islamabad) spanning banking, telecommunications, manufacturing, and IT sectors; time-lagged survey design; sector heterogeneity noted with manufacturing workers showing the largest negative effects. Themesgovernance human_ai_collab org_design IdentificationTime-lagged observational survey of employees analyzed with Partial Least Squares Structural Equation Modeling (PLS-SEM); tests of moderation (HR governance × AI managerial autonomy) and mediation (trust → commitment/satisfaction/turnover) using bootstrapped confidence intervals and an a priori power analysis; no random assignment or external instrumental variation, so identification is associational rather than causal. GeneralizabilityGeographically limited to Pakistan (three urban centers); cultural and regulatory context may not generalize to other countries., Sample drawn from 16 organizations — not a nationally representative or randomly sampled workforce., Sectors covered (banking, telecom, manufacturing, IT) but limited firm-level diversity and potential within-sector selection., Self-reported outcomes and perceptions may differ from objective productivity or turnover behavior., Observational design limits transportability of causal claims to other settings with different governance regimes or AI deployments.

Claims (13)

Claim	Direction	Outcome	Confidence & Evidence	Details
AI managerial autonomy is negatively associated with employee trust (β = –0.31, p < .001). Worker Satisfaction	negative	employee trust	Reading fidelity high Study strength medium	n=387 β = -0.31 0.3
AI managerial autonomy shows no significant association with organizational justice (β = –0.09). Worker Satisfaction	null_result	organizational justice	Reading fidelity high Study strength medium	n=387 β = -0.09 0.3
HR governance quality moderates the relationship between AI managerial autonomy and employee trust (interaction β = 0.29, p < .001). Worker Satisfaction	positive	moderation effect on employee trust	Reading fidelity high Study strength medium	n=387 interaction β = 0.29 0.3
HR governance quality does not moderate the relationship between AI managerial autonomy and organizational justice (β = 0.07, non-significant). Worker Satisfaction	null_result	moderation effect on organizational justice	Reading fidelity high Study strength medium	n=387 β = 0.07 0.3
Algorithmic transparency is the strongest HR governance predictor of employee trust (β = 0.44). Worker Satisfaction	positive	employee trust	Reading fidelity high Study strength medium	n=387 β = 0.44 0.3
Trust mediates HR governance effects on organizational commitment (indirect effect β = 0.31). Worker Satisfaction	positive	organizational commitment (indirect effect via trust)	Reading fidelity high Study strength medium	n=387 indirect β = 0.31 0.3
Trust mediates HR governance effects on job satisfaction (indirect effect β = 0.26). Worker Satisfaction	positive	job satisfaction (indirect effect via trust)	Reading fidelity high Study strength medium	n=387 indirect β = 0.26 0.3
Trust mediates HR governance effects on turnover intention (indirect effect β = –0.22). Turnover	negative	turnover intention (indirect effect via trust)	Reading fidelity high Study strength medium	n=387 indirect β = -0.22 0.3
Sector heterogeneity is observed: manufacturing employees are most negatively affected by AI managerial autonomy. Worker Satisfaction	negative	negative effect of AI managerial autonomy (presumably on trust and related outcomes) by sector	Reading fidelity medium Study strength medium	not reported 0.18
The study used a time-lagged survey of 387 employees from 16 organizations across Karachi, Lahore, and Islamabad in banking, telecom, manufacturing, and IT. Other	null_result	study sample and design (methodological fact)	Reading fidelity high Study strength high	n=387 0.5
PLS-SEM was used with bootstrapped confidence intervals and an a priori G*Power analysis confirmed adequate sample size. Other	null_result	statistical methods and power assurance	Reading fidelity high Study strength high	n=387 0.5
The paper advances the Human-AI Governance Framework (HAGF) by positioning HR governance as a moderating institutional mechanism, distinct from prior design-focused frameworks. Governance And Regulation	positive	conceptual advancement of a governance framework	Reading fidelity high Study strength speculative	not reported 0.05
Trust, not justice, is the primary casualty of ungoverned algorithmic management (i.e., ungoverned AMA reduces trust but not perceived organizational justice). Worker Satisfaction	negative	relative impact on employee trust versus organizational justice	Reading fidelity high Study strength medium	n=387 β_trust = -0.31 (p < .001); β_justice = -0.09 (non-significant) 0.3