Anticipatory guilt in chatbot service recovery

For a better understanding of the efficacy of chatbot‐based service recovery efforts, this study responds to calls to move beyond comparisons of recovery channels (e.g., human vs. AI agent) and examine how chatbots should act, especially relative to both economic recovery and emotional recovery, to enhance customer subjective well-being. Therefore, the present study tries to fill in this gap by specifically exploring the psychological mechanism for the effectiveness of different chatbot service recovery (SR) strategies. We review the literature on intelligent machines in service, SR, and prominent psychological theories of guilt. Then, we move on to explore anticipated guilt as a mechanism for moral intention. Next, we present our theory of guilt along with a series of propositions on AI recovery, and how it affects customer's subjective well-being. Previous research on emotion regulation strategies suggests that individuals faced with negative emotions often engage in one of two emotion regulation strategies: (1). Reappraisal, which involves re-evaluating a given situation to reduce or shift the negative emotion; (2). Suppression, which simply inhibits emotion-expressive behaviors. A specific emotional regulation strategy has distinct affective consequences, with reappraisal being more helpful than suppression at decreasing negative emotional experiences and promoting individual well-being (Haga et al. 2009). In negatively valenced situations such as addressing a service failure, emotional recovery strategy can alleviate negative emotions. Instead of suppressing negative emotions (focusing just on utilitarian SR), an AI agent providing emotional recovery shows empathic concern, encourages consumers to express their emotions through active listening and acknowledgement of such emotions. These actions can shift the consumers' perspective and facilitate the reappraisal of the situation (Groth and Grandey 2012). The ability of emotional chatbot recovery to regulate negative emotion should lead to subjective well-being conferred by forgiveness. Furthermore, extending on the theory of mind (Gray et al., 2007), we posit that AI agent with both human qualities, affective and cognitive capabilities, in negative encounter context could be perceived as human-like. AI agents appearing more humanlike often become more appealing and generate positive feelings. Individuals will apply and transpose their moral intention if they perceive AI to have more humanlike characteristics. In this research, we focus on guilt, a potential reaction of complaining customer in the private textbased service chats, as a moral emotion, a moral motivator for moral intention (Giroux et al., 2022; Kim et al., 2022). We elaborate a theory that proposes guilt as a negative emotion that increases subjective well-being in negatively valenced situations such as addressing a service failure, and provides insight into how it relates to an AI service recovery through a series of propositions. Interacting with economic and emotional chatbot SR would increase the customer's likelihood of engaging in forgiveness and enhance customer's subjective well-being through eliciting anticipatory guilt.