Explaining Loyalty to Generative AI: Extending the Theory of Planned Behavior with AI Anxiety and Perceived Risk
İbrahim Arpacı, Azize ŞAHİN, Azize ŞAHİN, Ekrem Tatoğlu, Aysun Sahin, Aysun Sahin · International Journal of Human-Computer Interaction · 2026
As generative artificial intelligence (GenAI) use grows, continuance intention does not necessarily imply loyalty. Drawing on the Theory of Planned Behavior (TPB), this study distinguishes continuance intention from loyalty, operationalized as affective attachment and personal significance, and examines AI anxiety and perceived risk as predictors of attitude. Competing deterrence and contextual salience expectations are specified for the risk–attitude relationship. Using PLS-SEM, we analyze data from 1,050 Generation Z users in Turkey in an AI-assisted travel-planning context. AI anxiety relates negatively to attitude, whereas perceived risk relates positively, supporting its predicted direction without establishing contextual salience as the underlying process. Both effects are modest and explain little variance in attitude. Attitude shows the strongest relationship with continuance intention among TPB predictors, while continuance intention is strongly linked to loyalty. The study advances GenAI post-adoption research by distinguishing continuance intention from loyalty and contrasting the roles of AI anxiety and perceived risk.