Perceived Response Failure of Mental Health Chatbots: Conceptualization and Consequences on Discontinuance

Artificial intelligence (AI) based chatbots offering mental healthcare services are getting much media attention recently (60 Minutes, 2024). Increasing adoption of these services is greatly facilitated by the significant shortage of mental health professionals to provide necessary services for an increasing number of patients who need these services. However, evidence suggests that chatbot users often complain about the insufficient capabilities of chatbots and discontinue use. To understand users’ perceptions of these insufficiencies, we introduce the concept of perceived response failure (PRF) to investigate associated factors and understand the impact of PRF on discontinuance of mental health chatbots (MHC) use. We use an ex post facto design to investigate our research hypotheses and expectation disconfirmation theory and social influence to explain the relationship between such PRF and use discontinuance.