Challenging AI Outputs: A Process Model of Distrust in Conversational AI
Abstract
While there is an increasing interest in examining trust in AI, distrust has received scant attention. This research focuses on distrust in conversational AI. Through analyzing interview data for ChatGPT users, we develop a process model of distrust in conversational AI. We show that distrusting conversational AI is a dynamic process that is triggered by various factors where users exercise agency and take an active role in addressing their distrust feelings through the mechanisms of confirmation, explanation, disambiguation, and efficacy. This results in either confirming or reversing their distrust feelings. Distrust, hence, can be momentary or enduring, and users adjust their interaction and use of the AI chatbot according to their experience. We discuss our findings and offer practical implications to the design of conversational AI.
Fields of Study & Subjects
Explainable Artificial Intelligence (XAI)
Ethics and Social Impacts of AI
AI in Service Interactions
Publication details
- Journal / source
- International Journal of Human-Computer Interaction
- Publisher
- Informa UK Limited
- Volume
- 42
- Issue
- 17
- Pages
- 14300–14313
- Publication date
- 12 January 2026
- Language
- EN