Artificial intelligence addiction among university students in China: risk stratification, consequences, and exercise-based intervention.

With the deep integration of artificial intelligence into daily life, AI-addiction-like tendencies have emerged as a potential form of behavioral dependency among university students. Focusing on high-frequency AI users on campus, this study examined risk stratification of problematic AI use, real-life impacts, and the preliminary pre-post improvements following exercise-based interventions. A cross-sectional survey assessed AI usage patterns and functional impacts, identifying a substantial proportion of students with at least mild risk of problematic AI use. Problematic AI use was associated with reduced academic performance, social engagement, physical health, personal interests, and emotional well-being. A six-month structured exercise intervention delivered to at-risk student was associated with reductions in problematic AI use risk and improvements in daily functioning. These findings provide preliminary support for conceptualizing AI addiction as a multidimensional behavioral dependency potentially distinguishable from traditional Internet addiction and suggest that structured exercise may hold potential as a non-pharmacological approach for mitigating maladaptive AI use in university settings.