AI-Enhanced Problem-Based Learning: Differential Effects of AI Usage Patterns on Core Competency Development Through Psychological Mechanisms

This study investigates how different levels of AI usage influence PBL core competencies through psychological mechanisms, establishing a theoretical framework that integrates Self-Determination Theory, Social Cognitive Theory, and Cognitive Load Theory. A cross-sectional survey design examined 412 university students across diverse academic disciplines. Participants completed validated instruments measuring AI usage patterns (assistive, dependent, substitutive), basic psychological need satisfaction, intrinsic motivation, self-efficacy, and five PBL core competencies (collaborative learning, autonomous learning, critical thinking, knowledge integration, communication skills). Structural equation modeling tested direct and mediated relationships. AI usage patterns demonstrated significant differential effects on PBL competencies. Assistive AI usage positively influenced all five core competencies, while substitutive AI usage negatively affected critical thinking abilities. Psychological mechanisms accounted for 64-73% of total effects through a sequential mediation pathway: basic need satisfaction → intrinsic motivation → self-efficacy → competency development. Competence need satisfaction emerged as the most critical mediator. The findings reveal that AI's educational impact depends critically on usage patterns and psychological mechanisms rather than technology presence alone. The study establishes theoretical foundations for AI-PBL implementation that prioritizes learner agency while leveraging technological affordances.