A recognizable gender bias on user’s satisfaction and intention to use healthcare chatbots

Recently, healthcare companies have delegated certain services offered to patients (e.g. the provision of information, the booking of medical visits, the payment of bills) to electronic conversation agents (chatbots) installed in virtual touchpoints. In order to make communications realistic, the designers endowed the communication interfaces with anthropomorphic characteristics such as faces, voices, and natural languages. In this study, the existence of a difference in terms of patient satisfaction and intention to use chatbots is assessed when the gender recognizable and attributable to the chatbot is male or female. Through an empirical study conducted on a sample of potential users, two stimuli (i.e., graphical interfaces of a chatbot) that simulated the booking of a medical examination were tested. The first stimulus was endowed with masculine characteristics and the second one with feminine characteristics. A mediation analysis showed a recognizable gender bias since a higher satisfaction was attributed to the stimulus with female characteristics that, in turn, led to a higher intention to use the service. Although this result has important theoretical and practical implications, further research should be conducted in order to identify the reasons and contexts in which this bias is most evident.