The Consumers Response to the Service Failure of Artificial Intelligent: The Role of Anticipatory Guilt

Interactions between humans and service robots are becoming a common part of daily life, but AI (Artificial Intelligent) service failures by companies are unavoidable. This research is focused on understanding how consumers react to failures of service robots and their likelihood of engaging in punishment behaviors as a result. Previous studies have shown that people perceive AI differently than humans in terms of agency and experience. Our four studies demonstrate that when faced with service failures, consumers are harsher with AI than with humans, and they increase their punishment behaviors because of reduced feelings of anticipatory guilt. We also show that the moderating effects of anthropomorphism and customers’ relationship orientation norms can reduce punishment behaviors. We have thoroughly ruled out alternative explanations for this effect, such as warmth, competence, and retaliation ability. Our findings offer insights into how consumers interact with AI failures and provide guidance for businesses in resolving AI service failures and developing AI services in business.