The Bright and Dark Sides of AI Dependency on Students’ Critical Thinking: The Roles of Cognitive Fatigue and Engagement

Abstract

Background/ problem: The rise of artificial intelligence (AI) use in education has sparked debate about its potential association with students’ critical thinking, while existing studies have reported inconsistent findings. Objective/ purpose: This study examined how AI dependency affects students’ critical thinking disposition through the mediating roles of cognitive fatigue and engagement. Design and Methodology: A cross-sectional survey was conducted among 639 university students in Java, Indonesia, using convenience sampling design. Data were collected through an online questionnaire and analyzed using partial least squares structural equation modeling (SEM-PLS). Results: The findings show that AI dependency does not directly affect students' critical thinking disposition (β = .05, p = .31). Instead, cognitive fatigue (β = -.60, p = .00) and engagement (β = .36, p = .00) significantly mediate the relationship in contrasting ways. On one hand, AI dependency increases cognitive fatigue, which subsequently weakens students' critical thinking disposition. On the other hand, AI dependency promotes engagement, which in turn strengthens critical thinking disposition. Conclusion and Implications: These findings suggest that AI dependency influence students’ critical thinking disposition indirectly through contrasting cognitive and behavioral pathways. From a behavioral-science perspective, modifiable student–AI interaction behaviors shape these outcomes. Educators should promote reflective AI engagement, require independent verification of AI-generated outputs, and design learning activities that maintain students’ responsibility for reasoning and decision-making. Strategies that limit cognitive offloading, uncritical acceptance of AI outputs, and excessive dependency help prevent cognitive fatigue and support students’ critical thinking disposition.