Do Perceived Privacy Risks of AI Matter? A Longitudinal Study on the Drivers of Continued Use of Intelligent Voice Assistants

Despite the rapid adoption of Intelligent Voice Assistants (IVAs), users often become passive in their continuous use. This study examines the possible drivers of IVAs' continuous use and the moderating effect of users' perceived privacy risk and purpose of use over a sustained period of use. Online survey data was collected from 818 respondents, over 12 months, using Qualtrics. The data was analyzed through structural equation modeling with SmartPLS 4.0. The results showed that over 12 months, the quality of interactivity, reliability, personalization, and satisfaction with the IVAs experience are key drivers of consumers' continuous use of IVAs while such influence is dependent on the purpose of use. However, IVA flexibility does not drive IVA's continuous use. The analysis showed that perceived privacy risk significantly moderated the effects of personalization and social presence. In contrast, no interaction effects were observed on the other paths examined. The research attempts to integrate social response theory and social presence theory that explain the humanlike attributes of IVAs (perceived interactivity, flexibility, reliability, personalization, satisfaction with the experience, social presence), and perceived privacy risk, to conceptualize IVAs' continuous use practices, to make a significant contribution to the IVAs literature.