Outsourcing thinking to AI? Focused immersion, AI dependency, and the double-edged impact on critical thinking

The rapid integration of artificial intelligence (AI) into higher education presents both cognitive opportunities and risks. Guided by a partial application of the I-PACE framework (operationalizing affective–cognitive reactions and executional tendencies), this study examines how perceived AI intelligence (PIAI) is associated with university students’ critical thinking (CT) and whether focused immersion and AI dependency mediate this association. Two questions are addressed: (a) Is PIAI directly associated with CT in AI-supported academic tasks? (b) Do focused immersion and AI dependency serve as mediators, including a sequential pathway? Cross-sectional survey data were collected from 698 Chinese undergraduates using validated scales. Confirmatory factor analysis and structural equation modeling with bias-corrected bootstrapped indirect effects were used to test the hypothesized relationships. Results indicate a positive direct association between PIAI and CT; focused immersion also shows a positive association with CT and operates as a positive indirect pathway. In contrast, a negative sequential pathway emerges whereby higher PIAI is linked to greater focused immersion, which increases AI dependency, and higher dependency is associated with lower CT. These findings delineate an attentional route (beneficial immersion) and a reliance route (detrimental dependency), clarifying when perceived capability supports versus undermines CT. The study extends I-PACE to educational-AI settings and informs course and policy design emphasizing verification-oriented engagement to sustain critical inquiry while mitigating reliance risks.