Abstract
As AI voice assistants such as Alexa become increasingly embedded in everyday routines, understanding how their design features shape user satisfaction and dissatisfaction is a growing concern for HCI and user experience (UX) research. Among these features, human-like voice is particularly salient yet insufficiently theorized. This study extends Herzberg’s Two-Factor Theory to human–AI interaction by examining how human-like voice influences users’ positive and negative evaluative experiences. Adopting a unipolar framework, we identify two mediating pathways, perceived usefulness and perceived enjoyment, through which human-like voice affects user evaluations. Results from an online survey reveal that enjoyment plays a more dominant role than usefulness, underscoring the affective core of AI user experience. This study advances our understanding of satisfaction and dissatisfaction as distinct but coexisting affective outcomes and offers design insights.
Recommended Citation
Yu, Yixiu; Souders, Dustin; Ning, Xue; and Davis, Fred, "Extending Herzberg’s Two-Factor Theory to AI Voice Assistants: How Human-like Voice Influences User Satisfaction and Dissatisfaction" (2025). SIGHCI 2025 Proceedings. 12.
https://aisel.aisnet.org/sighci2025/12