BAD ROBOT! BAD! PARALLELING DARK SIDE OF AI AND CONSUMERS’ MISUSE INTENTION: A FINANCIAL SCENARIO-BASED STUDY

Abstract

Artificial intelligence (AI) chatbots are increasingly deployed in financial services, yet their potential misuse by consumers remains underexplored. This study investigates how dark-side perceptions of AI-based banking chatbots influence consumers’ intention to misuse these systems. Drawing on cognitive, technological, and service-related perspectives, the study examines ten dark-side factors using a scenario-based survey of 393 postgraduate students. Data were analysed using covariance-based structural equation modelling (CB-SEM) and fuzzy-set qualitative comparative analysis (fsQCA). CB-SEM shows that sense of insecurity, distrust, perceived inauthenticity, service inequality, and service discrimination significantly increase misuse intention, whereas perceived risk, lack of familiarity, lack of personalization, confusion, and technophobia are not significant. The fsQCA results reveal multiple configurational pathways to misuse, demonstrating equifinality and highlighting the importance of interacting conditions. The findings extend AI dark-side research by conceptualizing misuse as a reactive behavioural response and emphasize trustworthy, transparent, secure, and fair chatbot design