With the evolution of Artificial intelligence (AI), emotional artificial intelligence service agents (AISA) have become common in service industry. However, how artificial empathy of AISA contributes to customer acceptance remains an open question. This study draws on Anthropomorphism Theory and Customer AI Experience Theory to examine whether and how artificial empathy has influence on customer acceptance of AISA. Evidence from three experiments (N=1057) designed by the Experimental vignette method (EVM) shows that: (1) artificial empathy including perspective-taking, empathic concern and emotional contagion has a positive impact on customer acceptance of AISA (study 1); (2) customer AI experience (emotional experience quality, social experience activation and social experience quality) mediates the relationship between artificial empathy and customer acceptance of AISA (study 2); (3) artificial empathy for hedonic (vs. utilitarian) services leads to a stronger effect on customer acceptance of AISA (study 3). This paper enriches our understanding of artificial empathy and provides practical guidance for practitioners strategically managing AISA services in AI-enabled marketing interactions.

Offering an apology is one of the common service failure recovery strategies. Previous studies focused on examining the effectiveness of apology from the customer perspective. It is not clear whether and how customers perceive firm remorse after an apology influence their blame attribution and coping behaviors. Integrating a cognitive-emotive model and an empathy model, this research proposes and empirically tests a remorse-empathy-coping model to explain how customers respond to apology after mobile application service failures occur. Specifically, this research examines how perceived firm remorse influences blame attribution and emotional empathy, which subsequently affects coping behaviors (revenge and avoidance) in the mobile app service recovery context. The moderating role of technology anxiety in the proposed model is also identified. Four hundred and fifty-two mobile application service users were recruited for a survey study and the Structural Equation Modeling was used in order to test the research hypotheses. Our findings show that perceived firm remorse negatively influences blame attribution but positively influences empathy. Empathy negatively affects revenge and avoidance behaviors. In addition, technology anxiety moderates the effect of perceived firm remorse on blame attribution. The negative effect of perceived firm remorse on blame attribution becomes weaker when technology anxiety increases.