What Makes An Apology More Effective? Exploring Anthropomorphism, Individual Differences, And Emotion In Human-Automation Trust Repair

Recent advances in technology have allowed an automation system to recognize its errors and repair trust more actively than ever. While previous research has called for further studies of different human factors and design features, their effect on human-automation trust repair scenarios remains unknown, especially concerning emotions. This paper seeks to fill such gaps by investigating the impact of anthropomorphism, users' individual differences, and emotional responses on human-automation trust repair. Our experiment manipulated various types of trust violations and apology messages with different emotionally expressive anthropomorphic cues. While no significant effect from the different apology representations was found, our participants displayed polarizing attitudes toward the anthropomorphic cues. We also found that (1). some personality traits, such as openness and conscientiousness, negatively correlate with the effectiveness of the apology messages, and (2). a person's emotional response toward a trust violation positively correlates with the effectiveness of the apology messages.