Unraveling the Influential Mechanisms of Smart Interactions on Stickiness Intention: A Privacy Calculus Perspective

Artificial intelligence (AI) technologies are changing the ways of interaction between humans and machines, and smart interactions have become one of the hot topics of artificial intelligent in-home voice assistants (AVAs) by connecting humans, machines, content, and AVAs. Based on the privacy calculus theory (PCT), the authors conducted an online questionnaire-based survey to investigate the influential mechanisms of smart interactions on stickiness intention (SI), demonstrated the positive (negative) effects of smart interactions on benefits and risks, and verified the moderating role of susceptibility to normative influence (SNI). The results show that smart interactions positively impact SI via utilitarian benefit and hedonic benefit; humanness has a U-shaped effect on privacy risk; personalization, connectivity, and linkage positively impact privacy risk; multimodal control negatively impacts privacy risk; and SNI positively moderates the effects of smart interactions on stickiness intention. The study enriched and expanded the literature on smart interactions in the context of AIoT and offered practical implications for AVA service providers and developers to design or optimize smart interactions for AI interactive services. By examining the double-edged sword effects of personalization and humanness, our findings offer novel insights into the privacy calculus in smart interactions.