AI Technology Adoption Leads to Better Performance: A Moderated Mediation Analysis of Self-Efficacy, Anxiety, and Resilience

Hing, Poh Hin (2026) AI Technology Adoption Leads to Better Performance: A Moderated Mediation Analysis of Self-Efficacy, Anxiety, and Resilience. Masters thesis, Tunku Abdul Rahman University of Management and Technology.

Text 35 Hing Poh Hin.pdf Restricted to Registered users only Download (5MB)

Abstract

The rapid integration of Artificial Intelligence (AI) into Malaysian workplaces has outpaced understanding of its psychological implications for employee performance. While organisations invest heavily in AI infrastructure, evidence suggests that technology adoption alone may not yield anticipated productivity gains. This study investigates the moderated mediation mechanisms through which AI Technology Adoption (AITA) influences Job Performance (JP), with AI Self-Efficacy (AISE) as a mediator and AI Anxiety (AIA) and Resilience (RES) as moderators among private sector employees in Malaysia. Grounded in Social Cognitive Theory, Conservation of Resources Theory, and Job Demands-Resources Theory, a quantitative cross-sectional design was employed. Data were collected from 362 eligible respondents via an online questionnaire and analysed using Partial Least Squares Structural Equation Modelling (PLS-SEM). Results indicated that AITA significantly predicted AISE (β = 0.181, p < 0.001), supporting the proposition that active AI engagement builds domain-specific confidence. However, neither AITA nor AISE demonstrated significant direct effects on JP, and the hypothesised mediation pathway was not supported. Moderation analyses revealed counterintuitive findings: RES negatively moderated the AITA→AISE and AITA→JP relationships, suggesting a resource substitution effect wherein highly resilient employees may rely less on technological resources for performance gains. AIA showed marginal positive moderation on the AITA→AISE pathway. Despite non-significant direct paths, the model demonstrated substantial explanatory power (R² > 0.90) and strong predictive relevance. These findings challenge linear assumptions in technology adoption literature and underscore the need for human-centric AI implementation strategies that address psychological safety, workflow integration, and complementary organisational support. The study contributes theoretical insights into boundary conditions of personal resources in digital transformation contexts and offers practical guidance for HR practitioners navigating AI adoption in emerging economies.