Hey There! What Do You Look Like? User Voice Switching and Interface Mirroring in Voice-Enabled Digital Assistants (VDAs)

We investigated user voice switching behavior (VSB) in voice-enabled digital assistants (VDAs), focusing on the importance of and preference for the voice accents, genders, and age to match with those of the users. We incorporated images of ten people with diverse races, ethnicities, age, genders, and religions to embody the voice interfaces (EVIs). In an online survey, we collected demographic, background, and VDA usage data. The sample consisted of 214 participants recruited through Amazon Mechanical Turk (http://mturk.com). The participants were selected based on owning a VDA (e.g., Alexa Home) or owning a device (e.g., smartphone, tablet, or computer), and setting the device on English as the default language. The age of the participants ranged from 18–35 years. Findings revealed that, regardless of age, the majority of the participants switched the voice interface and for various reasons. Further, participants placed importance on voice matching with their gender, accent, and age. Participants ranked the young White female, Asian female, and Black female EVIs as the most preferred for voice switching and interactions. We coin the concept, Interface mirroring, which should help designers to create more diverse and inclusive EVIs, ensuring fairness and equality in the design of VDAs.