Tomeito or Tomahto: Exploring consumer's accent and their engagement with artificially intelligent interactive voice assistants

Abstract

Artificially intelligent interactive voice assistants (AIIVAs) are developed to understand language, but there is limited insight into their ability to understand accents. While there have been substantial advancements in understanding multiple languages by AIIVAs, having an understanding of variety of accents is an emerging concern. To address these concerns, we contextualised our study in India, one of the world's most populated and diverse countries with varying accents and dialects. Study 1 collected qualitative data through semi structured interviews with participants, data was subsequently thematically analysed, and a typology was developed with respect to the context of use and consumers' emotional and rational reactions towards AIIVAs when interacting with accents. For Study 2, we implemented the quantitative research method. This was done to reiterate the conceptual model formulated from the qualitative research findings. Findings suggest that positive emotional action has emerged as the most significant factor, followed by rational action and negative emotional action. This study contributes significantly to the theoretical understanding of future consumer behaviour and human-computer interaction trends. It provides practical implications for managers, tech developers, and other companies working and using speech-to-text automatic speech recognition to know that while they train their algorithms with languages, they should be mindful of the diverse accents of their consumers.