Investigating Underrepresented Perceptions of Inclusively Designed Voiced Automation

Abstract

Consumers are increasingly interacting with automated systems for various purposes (e.g., smart speaker, vehicle control). One of the most popular ways they communicate with these systems is via the spoken auditory channel (listening and speaking). However, most of these systems tend to default to a female voice. In addition, many specifically use a White-sounding female voice. While prior work has investigated how attributes of age and gender influence perceptions of voiced automation, an additional vocal attribute of humanness has not been well researched -race. The specific goal of this project is to examine how more inclusively designed AI voices (e.g. exhibiting attributes of race) affect Black users’ attitudes toward autonomous technology. Results are expected to show more preferential attitudes when automation exhibits similar vocal attributes to the user. Findings will add to our understanding of how a deliberately more inclusive design affects an underrepresented population’s attitudes toward technology.