The challenge of AI dependence: exploring its impact on cognitive autonomy and elaboration in creative performance

Purpose As artificial intelligence advances and its application scenarios continue to expand, human–artificial intelligence interaction grows increasingly close. While artificial intelligence brings convenience, it also raises concerns that it may weaken users' cognitive autonomy. To investigate this issue. This study aims to explore the composition of AI dependence and its impact on cognitive autonomy and specific aspects of creative performance. Design/methodology/approach This study draws on cognitive evaluation theory and the elaboration likelihood model to construct a dual-pathway model with instrumental dependence and affective dependence as core variables. The model explains the formation mechanisms of the two types of artificial intelligence dependence and their differential effects on cognitive autonomy. It also preliminarily examines their potential role in the elaboration dimension of creative performance. A two-stage mixed-method design was employed to test the model. Findings Study 1 found that perceived trust mainly had a positive effect on instrumental dependence. Different dimensions of interactivity differentially affected affective dependence and cognitive autonomy. Mediation analysis further confirmed the core mediating role of Dependence. Study 2, a small-sample exploratory experiment, re-tested selected hypotheses and found a possible negative effect of instrumental dependence on the elaboration dimension of creative performance. Research limitations/implications The main theoretical contribution of this study lies in revealing the dual-pathway formation mechanisms of artificial intelligence dependence and the reshaping effects of different types of dependence on cognitive autonomy. It also preliminarily explores the potential effects of dependence on elaboration performance. Originality/value The dual-pathway dependence model constructed in this study offers practical implications for understanding how long-term use of artificial intelligence affects cognitive autonomy.