Interacting with voice-based artificial intelligence technologies: user perceptions of the dark side and the bright side

The nascent literature on voice assistants (VAs) investigates consumer perceptions by adopting partial theoretical and empirical perspectives. Given the fragmentation in the previous studies, our exploratory research aims to holistically explore the key dimensions of perceived risk and benefits related to user-VA interactions and to identify user clusters based on differences in perceptions. Quantitative exploratory research was conducted and the data were analysed with exploratory factor analysis (EFA) and k-means cluster analysis. The EFA results showed a three-factor solution: “data collection and misuse risk”, “utilitarian and hedonic benefits” and “symbolic benefits”. The k-means cluster analysis outlined three clusters: “iconic”, “rational and emotional” and “scared”. Our study offers the first holistic view of users’ positive and negative perceptions and provides practitioners with useful cues both in terms of interaction experience design and marketing communication.