AI ASSISTANCE, NOT AI DEPENDENCE, PREDICTS CRITICAL THINKING AMONG UNIVERSITY STUDENTS IN PAKISTAN

Abstract

Students use AI tools in two different ways: as a support for learning (AI Assistance) and as a replacement for their own effort (AI Dependence). This short article asks which of the two goes with stronger critical thinking. Data came from 150 university students in Pakistan (99 undergraduate, 51 postgraduate) who completed an online survey. AI Assistance had a moderate positive correlation with Critical Thinking (r = .522, p < .01) and AI Dependence a weak one (r = .173, p < .05). When both were entered together in a regression (R² = .272), only AI Assistance was a significant predictor (β = .517, p < .001); AI Dependence was not (β = .015, p = .843). The difference lies in how AI is used, not simply how much. Because the design is cross-sectional, the results show association and statistical prediction, not cause.