Exploring the ‘Dark Side’ of AI-Powered Digital Assistants: A Moderated Mediation Model of Antecedents and Outcomes of Perceived Creepiness

Abstract

Artificial intelligence-powered digital assistants (DAs) offer benefits, but their downsides can hinder user adoption and limit market potential. Although research exists on users' experiences with DAs, less is known about how user characteristics influence perceptions of a ‘dark side’—perceived creepiness. This study explores how perceived creepiness, user characteristics (uncertainty avoidance), and anxieties triggered by DAs (privacy concerns and technology anxiety) relate to distrust and disengagement from DAs. Based on the stressor-strain-outcome and person-environment fit models, a research model was proposed and tested using survey data from 509 DA users analysed with Amos version 29 software. The results reveal that privacy concerns, uncertainty avoidance, and technology anxiety positively influence perceived creepiness and distrust of DAs, leading to disengagement. Moreover, uncertainty avoidance moderates these relationships, with users higher in uncertainty avoidance experiencing a stronger link between privacy concerns/technology anxiety and perceived creepiness. Importantly, the study shows indirect effects mediated by perceived creepiness, with uncertainty avoidance further moderating these indirect relationships. These findings contribute to literature by highlighting the interplay between user characteristics, DA-induced anxieties, perceived creepiness, and user disengagement with DAs.