Generative Artificial Intelligence Dependency, Critical Thinking and Academic Decision-Making Behaviour Among Students in Polytechnics in South-West Nigeria

Abstract

Generative artificial intelligence is now embedded in students' academic practices, from writing to problem-solving. This study examined generative-AI dependency, critical thinking and academic decision-making behaviour among students in polytechnics in South-West Nigeria, and the relationships among them. A quantitative, survey-based correlational design was used, with a multistage sampling procedure yielding 240 students across six polytechnics. A validated questionnaire measured the constructs, and data were analysed using descriptive statistics, Pearson correlation, multiple regression and structural equation modelling. Generative-AI use was high (94.2%), with 68.3% using it daily or weekly for academic tasks; 52.5% reported difficulty completing tasks without AI and 49.2% reported uncritical acceptance of AI-generated content. Critical thinking disposition was moderate (mean = 3.18 on a 5-point scale), with truth-seeking (2.86) and open-mindedness (2.94) lowest. Decision-making behaviour was moderate (mean = 3.12). Dependency was negatively associated with critical thinking (r = −0.42, p < 0.001) and decision-making quality (r = −0.48, p < 0.001), while critical thinking was positively associated with decision-making quality (r = 0.54, p < 0.001). Critical thinking partially mediated the dependency effect on decision-making (indirect effect = −0.16, p < 0.01). The findings support AI-literate, thinking-centred pedagogy in Nigerian polytechnics.