Understanding evaluations of advice from chatbots : an exploration of advice response theory

• Main Authors: Chan, Share Yearn, Sebastian, Lydia, Sng, Shermaine, Zhang, Jia’En
• Other Authors: Andrew Prahl
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2021
• Subjects: Social sciences::Communication
• Online Access: https://hdl.handle.net/10356/147234
• Institution: Nanyang Technological University

This study extends Advice Response Theory (ART) to chatbot advisors and compares them to human advisors. We also explore how advice outcomes are affected when the element of uncertainty regarding advisor identity is introduced. Participants (N=443) rated sample advice messages from a human, chatbot or an “unknown” advisor, whereby the advisor could be either human or chatbot. We found that chatbot advisors were perceived less positively than humans which resulted in weaker influences of message content and worse evaluations of advice outcomes. Perceptions of unknown advisors were comparable with humans. Our findings provide insight for the continued use of ART in studying chatbot advisors and highlights the significance that the advisor identity has on advice evaluations.