Artificial intelligence (AI) has become an influential tool in higher education, shaping students’ learning behaviors and preparing them for future roles in industrial and information management. This study examines the dual impact of AI on student behavior, focusing on knowledge sharing and knowledge hiding, which are critical competencies in both academic and industrial contexts. Using conservation of resources (COR) theory and approach/avoidance orientation theory, we created a two-path model to investigate the impact of AI on proactive and restrictive behaviors among university students. Survey data from 501 Chinese university students revealed that AI usage enhances academic self-efficacy and knowledge sharing through approach-oriented behaviors while simultaneously increasing anxiety and knowledge hiding through avoidance-oriented behaviors. Additionally, academic demands and teacher support moderate these effects, suggesting practical strategies for AI integration in training environments. The findings offer valuable insights into managing knowledge sharing behaviors for future professionals in AI-driven industrial sectors, emphasizing the importance of balanced AI training to foster open collaboration and mitigate competitive knowledge concealment.