Who is to blame when AI fails? Attribution processes in anthropomorphic chatbot interactions

Abstract

The increasing integration of artificial intelligence (AI)–powered chatbots into digital customer service has fundamentally changed customer–brand interactions. While chatbots enable efficiency, accessibility and 24/7 service provision, their growing human-like design features also raise customer expectations regarding competence, responsibility and accountability. When chatbot failures occur, these heightened expectations may amplify customers’ emotional and behavioural reactions. Despite expanding research on chatbot performance and effectiveness, limited attention has been given to how customers cognitively evaluate and behaviourally respond to chatbot failures, particularly in relation to perceived anthropomorphism. Addressing this gap is essential for advancing service theory in AI-mediated contexts. [excerpt]