Thoughtless use of generative artificial intelligence and college students' self-directed learning: a multi-group SEM analysis of gender differences.

In recent years, generative artificial intelligence (GenAI) technologies have developed rapidly and have been increasingly integrated into educational settings. While students benefit from the convenience and efficiency provided by these tools, a growing tendency toward the thoughtless use of Generative Artificial Intelligence (TUGA) has emerged. However, the specific pathways and underlying mechanisms through which such thoughtless use influences undergraduates' self-directed learning (SDL) remain insufficiently examined. Grounded in Social Cognitive Theory (SCT), this study surveyed 487 undergraduate students from Henan Province, China, using a snowball sampling approach to collect self-reported data. Structural equation modeling (SEM) was employed to investigate the relationships among the TUGA, self-efficacy (SE), motivation (MOV), and SDL. The results demonstrate that the TUGA exerts a significant negative effect on undergraduates' SDL. Furthermore, SE and MOV partially mediate this relationship. Multi-group analysis further reveals substantial gender differences in the impact of the TUGA on SDL, MOV, and SE. By elucidating the internal mechanisms through which the TUGA undermines undergraduates' SDL, this study provides both theoretical and empirical contributions, offering practical implications for higher education institutions seeking to promote responsible AI use and strengthen students' SDL.