Psychotherapy chatbot design proposal: Empirical tests of user’s psychological responses and self-disclosure to chatbots with personalities

The current research demonstrates empirical tests of user’s affective responses and self-disclosure to text-based psychotherapy chatbots with personalities. Experiment 1 will identify how people would respond to chatbots that are either dominant or submissive. It will be conducted by a 2 (chatbot personality: dominant vs. submissive) X 2 (subject personality: dominant vs. submissive) between-subject experiment. Experiment 2 will provide evidence that intimate information disclosures influence how people behave in subsequent interactions with a chatbot. It will be conducted by a 2 (question-wording: reciprocal vs. non-reciprocal) X 2 (interview sequence: gradually vs. abrupt) between-subject experiment. I expect people will be more affected by chatbots with personalities, and will be more attracted to psychologically similar chatbots to themselves. In addition, when people are exposed to intimate information exchanges through the gradual conversation with chatbots, they will be more likely to disclose themselves to chatbots in turn.