Voice-based personal assistant (VPA) trust: Investigating competence and integrity

Abstract

While intelligence of voice-based personal assistants (VPAs) is evolving, limited research has investigated how various intelligence attributes impact consumer trust and adoption of VPAs. Drawn from social agency theory and the technology acceptance model, the current study developed and empirically tested a holistic research framework of the effects of intelligence attributes on consumers’ behavioral responses to VPAs through trust. We identified three aspects of intelligence attributes as functional (ease of use, usefulness), hedonic (enjoyment, innovativeness), and anthropomorphic attributes (humanness, autonomy) and two types of trust in VPAs as competence and integrity. Results showed dynamic relationships among VPAs’ intelligence attributes, consumer trust, and behavioral intentions. Specifically, both functional and hedonic attributes promote competence, whereas anthropomorphic attributes increase integrity. Further, both competence and integrity lead to positive word-of-mouth (WOM), while only competence positively influences consumers’ usage intention. In addition, a user's level of technology anxiety (an individual difference variable) was found to serve as a moderator in the trust model. This study offers both theoretical contributions and practical implications regarding consumer trust development in the context of VPAs.