Designing for Inclusion: Exploring How the Accent, Tone and Gender of an Artificially Intelligent Voice Affects Usability

Abstract

Communication is a two-way street. There has been research and work done on the focus of how language is interpreted as it is delivered from the user to the Artificially Intelligent applications and products, but what about the intricacies of how the machine responds back? Is a familiar accent and particular tone response from the application and or product important? Does the gender of the Artificially Intelligent agent foster a sense of comfort for the user? Looking further, what are the potential or necessary design guidelines that could envision a future with more inclusive vocal output by Artificially Intelligent devices and systems. An online survey consisting of both qualitative and quantitative items was administered to twenty participants (N = 20), and the resulting data were analyzed. Across this sample set, participants (N=20) indicated that usability was largely affected by tone in relation to the gender and accent of the vocal output. Results within the examined group also indicate a lower receptivity among participants(N=20) to a gender-neutral Artificial Intelligence (AI) voice option. Guidelines and best practices were highlighted to aid future designers in incorporating more nuanced voices within the field of Artificial Intelligence.