The dual role of ChatGPT in shaping postgraduate creative outcomes: Evidence from Vietnam

Abstract

This study investigates the dual effects of ChatGPT usage on postgraduate students’ creative outcomes within Vietnamese higher education. Based on a quantitative survey (N = 312) and analyzed using Partial Least Squares Structural Equation Modeling (PLS-SEM), the results show that ChatGPT usage positively influences creative self-efficacy (β = 0.42, p < 0.001) while negatively impacting original thinking (β = -0.30, p < 0.001). Perceived usefulness significantly moderates these relationships, strengthening both the positive (β = 0.18, p = 0.004) and negative (β = -0.15, p = 0.019) effects. Multigroup analysis further reveals meaningful differences across usage frequency, technological familiarity, and academic disciplines, with high-frequency users experiencing a greater trade-off between enhanced creativity and diminished originality. These findings advance AI-in-education literature by highlighting ChatGPT’s complex cognitive impacts in emerging economies and refine creativity theories by demonstrating how perceived AI utility shapes creative outcomes. Practical implications include targeted AI literacy initiatives and differentiated integration strategies to maximize benefits while mitigating cognitive risks. Future research should address limitations related to cross-sectional design and expand across broader cultural and technological contexts.