Original Research

Algorithmic eyes, exit minds: A boundary–autonomy cascade model of artificial intelligence monitoring and turnover intention among Indonesian Generation Z employees

Andi Adawiah, Nurmal Idrus
SA Journal of Human Resource Management | Vol 24 | a3699 | DOI: https://doi.org/10.4102/sajhrm.v24i0.3699 | © 2026 Andi Adawiah, Nurmal Idrus | This work is licensed under CC Attribution 4.0
Submitted: 08 March 2026 | Published: 14 September 2026

About the author(s)

Andi Adawiah, Department of Management, Universitas Lamappapoleonro, Soppeng, Indonesia
Nurmal Idrus, Department of Management, Universitas Lamappapoleonro, Soppeng, Indonesia

Abstract

Orientation: Artificial intelligence (AI) monitoring differs from earlier electronic performance monitoring in three respects: It observes continuously, it infers states that employees never disclose, and it removes human discretion from evaluation. These features change how employees experience oversight.
Research purpose: We examine how perceived AI monitoring relates to turnover intention (TI) through two sequential mechanisms: Perceived privacy intrusion (PPI) and psychological reactance.
Motivation for the study: Monitoring research disagrees about whether oversight harms employees, yet it rarely specifies the psychological route from algorithmic oversight to exit cognition. We address that route directly rather than restating the debate.
Research approach/design and method: We collected cross-sectional survey data from 390 Generation Z employees in Indonesian organisations whose work is subject to automated monitoring. We tested the model using partial least squares structural equation modelling (PLS-SEM).
Main findings: Artificial intelligence monitoring relates most strongly to PPI. Intrusion predicts psychological reactance, and reactance is the most proximal predictor of TI. The central mechanism is a serial cascade: Monitoring raises intrusion, intrusion activates reactance, and reactance drives the intention to leave.
Practical/managerial implications: Because reactance, not intrusion, is the proximal driver of exit, technical privacy fixes alone are insufficient. Restoring discretion, voice and procedural justification addresses the mechanism that the data identify as decisive.
Contribution/value-add: We specify a boundary–autonomy cascade. The contribution is the ordering of appraisal before motivation and before behaviour, not the integration of established theories as such. This ordering reframes turnover as an autonomy-restoration response to oversight that employees judge illegitimate.


Keywords

AI monitoring; privacy intrusion; psychological reactance; turnover intention; algorithmic management; workplace surveillance.

JEL Codes

J28: Safety • Job Satisfaction • Related Public Policy

Sustainable Development Goal

Goal 8: Decent work and economic growth

Metrics

Total abstract views: 1171
Total article views: 1993


Crossref Citations

No related citations found.