Positive Affect and Academic Skill Development Through ChatGPT in Higher Education

This study investigates how positive affect during ChatGPT-4 use is associated with perceived academic skill development in higher education, emphasising the mediating role of attitudes toward ChatGPT and the moderating role of learning orientation, grounded in Broaden-and-Build Theory and the Technology Acceptance Model. The enrichment of academic skills through ChatGPT is proposed to be maximised when students who experience positive affect develop positive attitudes toward the tool and channel their interactions through deep learning orientations. This study uses quantitative analysis with n = 12,035 active ChatGPT-using students from 135 countries. PLS-SEM using bootstrapping with 200 resamples is employed for data analysis. The results confirm that positive affect is positively associated with perceived academic skill development (H1: β = 0.199, p < 0.001) and attitude (H2: β = 0.267, p < 0.001), and that attitude is significantly associated with skill development (H3: β = 0.354, p < 0.001). Attitude partially mediates the positive affect to skill development relationship (H4: β = 0.095, t = 24.923, p < 0.001). Learning orientation negatively moderates both the attitude skill development (H5: β = −0.021) and positive affect skill development (H6: β = −0.045) pathways, indicating a desirable difficulty effect: when learning orientation is high, less evaluative scaffolding is needed for skill acquisition, consistent with cognitive challenge theory. What this implies for practice is that access alone is not enough. Institutions need to cultivate AI interactions that feel emotionally positive and, alongside them, a learning-oriented form of engagement; only when the two are developed together does ChatGPT’s potential for academic skill development carry across the diverse global contexts of higher education.