Talking with AI about Mental Health: The Role of Perceived Message Contingency and Outcome Relevant Involvement

Kim, H. (Grady Alum), Zhang, W., Han, J. Y., Seo, J. K. (Ph.D. candidate), Yoon, H. J., & Yeon, J. (Ph.D. student) (2024, June). Talking with AI about Mental Health: The Role of Perceived Message Contingency and Outcome Relevant Involvement. Paper to be presented to Human-Machine Communication Division of the International Communication Association Annual Conference, Gold Coast, Australia.

Abstract: Recognizing the growing utilization of conversational artificial intelligence (AI) technologies in everyday communication, this study explored the impact of interacting with AI chatbots on mental health outcomes. Using data from a field experiment (N = 316) involving the use of ChatGPT, this study showed that engaging in conversations with AI chatbots about stressful issues, in addition to utilizing online health resources, increased levels of perception of message contingency when compared to using online health resources alone. Consequently, perceived message contingency had a positive effect on participants’ subjective well-being through cognitive and affective adjustments. Notably, the influence of AI chatbot use on perceived message contingency was more pronounced among participants who attached greater personal importance to a stressful issue. Theoretical and practical implications are discussed.