Managing turn-taking through beep sounds by Bixby: Applying conversation analysis to human-chatbot interaction

초록

This paper explores turn-taking between users and Bixby, a Samsung voice assistant. While using a simplified turn-taking system designed to guide users, Bixby does not fully understand and replicate the complex turn-taking system that characterizes human interaction. Utilizing Conversation Analysis (CA), this paper highlights the limitations of Bixby’s turn-taking system, which relies on mechanical cues such as ding↓↑ to indicate turn initiation and turn completion. These cues prove inadequate in accommodating the dynamic nature of human interaction, causing confusion in users regarding when and how to talk to the digital assistant. Users often predict turn endings based on their familiarity with human turn-taking, which is often incompatible with Bixby’s rigid turn-taking mechanisms. This study incorporates insights form CA and applies sequential analysis to interactions between humans and the voice digital assistant to explicate how Bixby’s turn-taking mechanisms become problematic. This information helps explore ways to improve these interactions. © 2023 Ji Young Lah · Yo-An Lee, published by Linguistic Research (KHU ISLI). This work is licensed under the Creative Commons Attribution 4.0 International License.

키워드