Explainable AI in Digital Banking: Customer Responses to Technological Risk Through Algorithmic Trust and AI Anxiety

Artificial intelligence (AI) features in FinTech-enabled digital banking support financial decisions but also introduce technological risks, such as black-box reasoning and erroneous recommendations. Studies often treat customer-facing AI as a single construct, leaving it unclear whether customers respond to explanation consistently across features. Drawing on the Stimulus–Organism–Response framework, this study models perceived AI explainability as a stimulus related to continuance intention through algorithmic trust (a cognitive pathway) and AI anxiety (an affective pathway) and compares chatbot and robo-advisor users. Survey data from 300 Indonesian digital banking users (160 chatbot users; 140 users of self-reported financial recommendation or automation features, labeled robo-advisor users) were analyzed with partial least squares structural equation modeling and multi-group analysis. Explainability is associated with higher continuance intention directly and through stronger trust and lower anxiety; technological risk serves as the theoretical context, and because neither actual risk nor perceived risk was measured, the findings do not show that explanation reduces risk. Neither indirect association differs significantly between self-selected groups. At the path level, anxiety is more strongly associated with lower continuance among robo-advisor users, and explainability with lower anxiety among chatbot users, the latter not being robust to multiple-comparison adjustment. The findings inform customer-side research on AI in a technological risk context and motivate, but do not test, a common explanation standard complemented by human oversight for financial recommendation features.