Generative AI Literacy and Critical Thinking among University Students: The Mediating Role of AI Verification Practices

Abstract

Data were gathered through a quantitative, cross-sectional survey of around 400 undergraduate and postgraduate students from Islamabad universities in Pakistan, with stratified random sampling used to cover different academic fields. We captured generative AI literacy across four dimensions (understanding of AI capabilities, prompt literacy, awareness of AI limitations, and ethical awareness), gauged AI verification practices by how often students fact-checked, cross-checked sources, weighed AI-generated material, and spotted hallucinated or misleading text, and tracked critical thinking with a validated critical-thinking scale. The data were evaluated through descriptive statistics, reliability checks, correlation analysis, hierarchical regression, and bootstrapped mediation analysis. These initial results show that generative AI literacy tracks positively with critical thinking, that AI verification practices link positively to both, and that verification practices partially mediate between literacy and critical thinking. These results suggest that literacy on its own cannot shield or sharpen critical thinking in environments dense with AI; rather, regularly checking AI outputs acts as a primary route through which literacy builds better critical thinking. The paper concludes by discussing what this means for curriculum design, digital literacy initiatives, and academic integrity policies in higher education across Pakistan.