Background - AI conversational agents have demonstrated efficacy in social contact interventions for stigma reduction at a low cost. However, the underlying mechanisms of how interaction designs contribute to these effects remain unclear.
Our Work - This study investigates how participating in three human-chatbot interactions affects attitudes toward mental illness. We developed three chatbots capable of engaging in either one-way information dissemination from chatbot to a human or two-way cooperation where the chatbot and a human exchange thoughts and work together on a cooperation task. We then conducted a two-week mixed-methods study to investigate variations over time and across different group memberships.
Findings - The results show that, compared to the one-way chatbot, interacting with one of the cooperation task chatbots was associated with perceiving the chatbot to be more competent and likeable, and with expressing more empathy during the conversation. Additionally, the results indicate that the interaction mode significantly influenced participants’ perceived relationships with chatbots and attitudes toward mental illness. While human-AI cooperation demonstrated effective stigma reduction, concerns arose regarding the lack of inconsistency between agents' task performance and metaphor. We discuss the implications of our findings for human-chatbot interaction designs aimed at changing human attitudes.
Keywords: Chatbots; Conversational Agents; Social Stigma; Mental Illness
Authors: Tianqi Song, Jack Jamieson, Tianwen Zhu, Naomi Yamashita, Yi-Chieh Lee