How does AI technological uncertainty affect continuance usage intention of Voice AI Assistants? A valance-based framework

Purpose Voice artificial intelligence (AI) assistants have gained widespread usage, uncertainty surrounding their adoption remains prevalent (e.g. unknown safety, technological limitations, service limitations, and lack of humanization). Therefore, our study aims to better understand the correlations among elements that influence the continuance usage intention of Voice AI Assistants by investigating the impact of AI technological uncertainty on perceived utilities of consumers and services. Design/methodology/approach Based on the expectation theory and stimulus-organism-response (SOR) model, we develop a research model and collect data via a survey. Then we adopt structural equation modeling and fuzzy-set qualitative comparative analysis (SEM-fsQCA) method to examine the research model. Findings The findings reveal that AI technological uncertainty influence users’ high and low continuance usage intention of Voice AI Assistants through two distinct but interconnected pathways. The first path highlights the increase in negative utilities, which further enhance the low continuance usage intention of Voice AI Assistants. The other path is to increase the positive utilities, which further enhances the high continuance usage intention of Voice AI Assistants. Moreover, the findings show that uncertainty, positive utilities and negative utilities have configurational effects on users’ continuance usage intention of Voice AI Assistants. Originality/value This study makes a distinctive contribution by integrating both technological uncertainty and perceived utilities—both positive and negative—into a comprehensive model that better explains the complexities of consumer continuance usage intention in the context of Voice AI Assistants. Furthermore, the results highlight the configurational effects of these factors, presenting a comprehensive perspective on how various configurations of uncertainty and utilities shape user intentions.