Understanding ChatGPT adoption in higher education: An affective-cognitive dual-pathway model

This study develops and tests a dual-pathway model of ChatGPT adoption in higher education, in which positive emotions support adoption through perceived capabilities, performance expectancy and satisfaction, while negative emotions constrain it through ethical concerns. The analysis reuses secondary data from a global survey of higher education students, collected through non-probability convenience sampling with voluntary participation. After applying the eligibility criteria, the final sample comprises 11,451 ChatGPT users from 90 countries. Covariance-based structural equation modelling and multigroup analysis were used. The two pathways are strongly asymmetric. The total indirect effect of positive emotions on intention was 4.5 times larger than that of negative emotions, and performance expectancy was the dominant predictor of intention. Ethical concerns reduced intention significantly but negligibly. Four of the ten paths differed by usage intensity. The study integrates affective and cognitive explanations of adoption within one framework and supports emotionally informed strategies for AI integration.