The Impact of Robot Reliability on User Affection: A Moderated Serial Mediation Analysis
Ruo-Yan Wu, Hui Zhi Lin, Ying Huang, Dan Chen, Zhentao Liu, Chen-Ling Liu · International Journal of Human-Computer Interaction · 2025
The integration of robots into daily and professional life is becoming increasingly widespread. Consequently, users’ expectations of mechanical perfection are shifting towards perceiving robots as human-like interaction partners. Given the dynamic cognitive processes involved in human-robot interaction, whether a robot’s errors reduce user affection remains a subject of ongoing debate. This study incorporates self-efficacy and algorithmic attitudes as mediating variables, a serial mediation model is proposed to explore the impact of robot reliability on user affection, with robot anthropomorphism as a moderating factor. Empirical analysis of 622 valid responses reveals that: self-efficacy positively influences algorithmic attitudes, forming a serial mediation effect between robot reliability and affection; anthropomorphism negatively moderates the relationship between algorithmic attitudes and user affection. The research results provide a new perspective for studying the complex psychological processes of human-computer interaction and offer insights for the development and design of advanced interaction systems.