Effect of AI Chatbots´ Gender in the Business Context: Empirical Evidence from Debt Collection

Abstract

Chatbots have been empowered by Artificial Intelligence (AI) and rapidly applied to many industries. There is a call for more understanding of the effect of chatbots’ social cues on business outcomes. This paper investigates how does the choice of chatbots’ voice gender impact customer’s intention to repay overdue debt. Prior studies on gender differences have conflicting implications. Employing unique real business dataset, we find that for male customers, they are more willing to repay when served by female chatbots. However, female customers have no preference for the gender of chatbots. We explain the mechanism of gender effect by finding that female customers lead to more engagement of male customers, hence induce higher repayment intention. We also find that the advantage of female chatbots are more prominent when the calls happen in the evening (when people have more emotional needs) and when the customers are less likely to repay initially. The main effect is validated by several robustness checks. Especially, this paper finds that the gender-stereotypical attributes (e.g., forceful and assertive of masculinity, gentle and warm of femininity) have the same effects as chatbot gender. We further discuss the theoretical contributions and managerial implications.