Designing AI for Human-Human Interaction
November 10 @ 12:00 pm - 1:00 pm
Rokwire Community Lecture Series
Lecture by Yun Huang, School of Information Sciences (iSchool)
Attend in person or virtually. Complete this form to receive a zoom link.
Chatbots are a promising technology for delivering services for multiple applications and across many domains. Specifically, prior studies have shown that chatbots have the potential to coach users who are learning new and different skills. However, several limitations of chatbot-based approaches remain. People may become disengaged from using chatbot-guided systems and fail to follow the related guidance.
In this talk, I will introduce our research on designing and evaluating conversational AI that delivers guidance for people practicing journaling skills. I will present a series of studies on novel chatbot designs that improve human-chatbot interaction, users’ perceived engagement and trust, and the long-term effects of using chatbots for practicing new skills.
Our findings show that users engage more effectively when chatbots incorporate human experts’ guidance, even though users could perceive higher pressure and feel less motivated to continue the practices voluntarily. Our work also suggests reciprocity can occur in human-chatbot interaction, i.e., when a chatbot self-discloses more like a human, users also disclose their feelings and thoughts more deeply to the chatbot with more trust.