Presentation
Visual Attention in Healthcare Chatbot Interactions: Quantifying Varying Communication Styles with K-Coefficient
DescriptionAdvancements in artificial intelligence (AI) are transforming healthcare chatbots, improving their potential to support patient education and engagement. The effectiveness of healthcare chatbots depends not only on technical ability but also on how their communication style impacts user engagement, particularly through visual attention. To investigate this, a between-subjects study evaluated the effect of chatbot communication style (conversational vs. informative) on eye-gaze behavior in a knowledge-seeking task. Eye metrics were analyzed using quantile and linear regression models. Quantile regression models revealed the distribution of fixation duration was always higher in the informative condition, while saccadic amplitude and the K-coefficient varied across quantiles. The linear regression model showed that both communication style and stimulus progression significantly increased the K-coefficient over time. These findings demonstrate that chatbot communication style influences user visual attention over time, underscoring the need for future work to align chatbot communication styles with user attention patterns.
Event Type
Industry/Practitioner Content
Lecture
TimeTuesday, October 14th11:30am - 11:50am CDT
LocationGrand A
Cognitive Engineering & Decision Making
Similar Presentations