Presentation
The Effects of Citations and Confirmation Bias on Trust in Chatbots
SessionPoster Session 1
DescriptionConversational AI chatbots based on large-language models offer promising applications in commercial, educational, and personal use, but only if they are accepted. To increase trust, many chatbots provide users with citations as a form of explanation and social proof, but prior literature is unclear about whether citations remain effective in elevating trust when response content contradicts a user’s prior beliefs. A 3x2 within-subjects survey asked 100 users their trust in conservative, moderate, and liberal responses to political questions, with and without citations. Responses were categorized as confirming or contradicting existing beliefs based on self-reported political lean. A linear mixed-effects model showed that users have significantly higher trust in responses when responses are moderate or confirm beliefs, but that citations do not significantly affect trust. These results suggest that citations may not be effective when addressing politically controversial topics, and that balanced responses may more effectively garner user trust.
Event Type
Poster
TimeTuesday, October 14th5:30pm - 6:30pm CDT
LocationRiverside East