VUI-Speak: There Is Nothing Conversational about “Conversational User Interfaces”

Abstract

In this chapter, we suggest a concept for describing participants' practices regarding progressively adapting their actions to fit the computational system in voice user interfaces (VUIs) such as Google Home.We describe this phenomenon as "VUI-speak."Although developers aim at enabling computers to communicate like humans, our study shows that, on the contrary, people accommodate the device through VUI-speak.Based on video ethnographic studies and ethnomethodological conversation analysis (EM/CA) of blind people's natural use of Google Home, this research contributes to EM/CA studies of human-computer interaction, human-robot interaction, and VUIs in particular.The research findings suggest (1) that VUIspeak is produced at the third position in a five-part sequential structure, (2) that a change in action formation occurs, and (3) that this change relates to producing what we call an "application-oriented turn."This research has practical implications for the design of conversational systems and contributes to the expanding field of EM/CA research on VUI interaction.