The association between interpersonal alienation and problematic use of conversational AI among adolescents: A mediated network analysis of perceived anthropomorphism and AI attachment

Background and aims: With the rapid development of conversational artificial intelligence (CAI), problematic use of conversational AI (PUCAI) among adolescents has attracted increasing attention. However, the underlying psychological factors associated with PUCAI remain unclear. Guided by the Interaction of Person-Affect-Cognition-Execution (I-PACE) model, this study explored the associations between interpersonal alienation, perceived anthropomorphism, AI attachment, and PUCAI. Methods: Data were collected from 1,124 Chinese high school students (516 males, 608 females; Mage = 16.59). Serial mediation analysis and network analysis were conducted. Results: Interpersonal alienation was positively associated with PUCAI. It also had indirect associations through both parallel and serial mediation involving perceived anthropomorphism and AI attachment. Network analysis further revealed that perceived anthropomorphism exhibited the highest bridge expected influence, and emotional trust showed the highest centrality. Conclusion: These findings suggest that adolescent PUCAI is associated not only with general psychological risk factors but also with social-cognitive perceptions and emotional bonding experiences during human-AI interactions. This study provides a novel perspective for understanding problematic use behaviors in emerging human-AI interaction contexts and offers implications for future risk assessment and research on healthy AI use.