You Go First: The Effects of Self-Disclosure Reciprocity in Human-Chatbot Interactions

Chatbots have several affordances that may stimulate self-disclosure and help-seeking by people with mental health problems, such as accessibility, anonymity, and convenience. This study aims to examine the effect of reciprocal self-disclosure by chatbots on friendship formation between humans and chatbots. For this, a Wizard of Oz experiment with two conditions was conducted. Participants (N = 86) were randomly assigned to either the condition with or without reciprocal self-disclosure by the chatbot. Our preliminary findings revealed that there was no significant difference between the conditions regarding feelings of friendship. However, the results showed that people trusted the chatbot in the reciprocal self-disclosure condition more. Furthermore, people perceived the chatbot in the reciprocal self-disclosure condition as more empathetic. These first results suggest that in developing affective social chatbots it would be important to implement reciprocal self-disclosure as it stimulates perceptions of trust and empathy.