Let’s Work Together: Integrating Human Support with Conversational Agents

Yi-Chieh Lee; Naomi Yamashita; Yun Huang

Let’s Work Together: Integrating Human Support with Conversational Agents

Yi-Chieh Lee ,
Naomi Yamashita ,
Yun Huang

August 2020

Download BibTex

ABSTRACT

Chatbots have the potential of eliciting users’ deep self-disclosure and mediating human-to-human interaction. However, little is known about how people would interact differently with a human-supported chatbot than when talking to a chatbot alone. We design chatbots with (CH) and without (CO) human support to deliver suggestions for people to practice journaling skills. We conduct a study to investigate the effects of the two chatbot designs and each participant used their chatbot for four weeks. Our results show that the CH participants perceived a higher level of engagement than the CO participants when they received suggestions for practicing journaling skills and also engaged in deeper self-disclosure. However, after finishing the journaling-skill training session, the CO participants were more willing to keep practicing the suggested journaling skills than the CH group. The COVID-19 pandemic has heightened the challenges of providing healthcare services and social connection. Our in-progress research proposes a human-support chatbot system to explore effective designs to facilitate users’ social connection and well-being and through conversational user interfaces.

Keywords

chatbot, self-disclosure, human support, conversational user interfaces

ABOUT THE AUTHOR/S

YI-CHIEH LEE
University of Illinois at Urbana-Champaign
ylee267@illinois.edu

Yi-Chieh Lee is a doctoral student at Computer Science. His research interests include human-computer interaction (HCI), social computing, and AI & HCI. He has published in top tier venues, e.g., CHI, CSCW, IUI, and several Journals. Currently, he is working under Prof. Yun Huang to finish his dissertation, which is about designing effective conversational agents to promote mental wellbeing and behavior change. His PhD thesis is sponsored by NTT research Lab.

NAOMI YAMASHITA
NTT Communication Science Laboratories
naomiy@acm.org

Naomi Yamashita is a Primary Researcher at NTT Communication Science Laboratories. Her primary interests lie in the areas of computer-supported cooperative work and computer mediated communication. Using a combination of quantitative and qualitative research methods, she aims to uncover the nature of human discourse/interaction and to propose guidelines for designing new communication technologies. Her current projects focus on the design, development and evaluation of technologies for mindful inclusion in various domains such as global teams and mental healthcare.

YUN HUANG
University of Illinois at Urbana-Champaign
yunhuang@illinois.edu

Yun Huang is an assistant professor in the School of Information Sciences at the University of Illinois at Urbana-Champaign. She co-directs the Social Computing Systems (SALT) Lab. Before joining Illinois, she was a faculty member in the School of Information Studies at Syracuse University and a postdoc fellow at Carnegie Mellon University. Her work focuses on social computing systems research, in which she examines context-driven approaches of designing crowdsourcing systems. For example, her research examines how different contexts impact user contributions to crowdsourcing systems; how to leverage these contextual effects to design innovative social computing systems that can better engage users; and how crowdsourced user behavioral data may help better understand users and the community. She received her PhD from the Donald Bren School of Information and Computer Sciences at the University of California, Irvine.