OVERVIEW
The rapid emergence of AI-enabled “digital humans,” commonly referred to as virtual influencers, introduces a new class of programmable digital artifacts that mediate interactions between organizations and consumers. Although prior research has examined influencer marketing effectiveness, limited attention has been given to understanding virtual influencers as information systems characterized by distinct design attributes and interactional qualities. Drawing on the Information Systems Success Model and the Elaboration Likelihood Model, this study conceptualizes virtual influencers as AI-driven artifacts whose system quality, information quality, and interactional properties shape user cognition and relational engagement. Using large-scale survey data, we examine how AI artifact qualities influence perceived social presence and parasocial interaction, and how these mechanisms affect luxury brand attitudes, engagement intentions, and purchase intentions. By integrating persuasion theory with IS quality dimensions, the study advances a process model explaining how users evaluate and respond to AI-mediated communication systems in high-involvement consumption contexts. This research contributes to IS scholarship by extending digital artifact theory to AI-generated personas, clarifying how system and information qualities translate into relational outcomes, and offering insights into the design and governance of AI-enabled digital agents in organizational settings.
The Information Privacy
Virtual Influencers as AI-Enabled Digital Artifacts: A Theory-Driven Examination of Audience Perceptions Soo Il Shin
A Multilevel Process View
In the digital age, information is both a vital means for progress and a crucial asset to protect. Society, organizations, and individuals differ greatly in how they balance the benefits of information use with the rights and ability to protect personal information, creating a wicked social justice problem. We propose the concept of the information privacy divide, which refers to systematic differences in information privacy access, capabilities, and outcomes across societal, organizational, and individual levels. We develop a multilevel, process-oriented framework that examines how society’s privacy regulation cascades down to influence individual perceptions and behaviors, which, in turn, produce outcomes at all three levels. We test this framework using survey data from representative samples in three countries with markedly different privacy regimes. This study contributes to the information privacy literature by introducing a holistic, multilevel perspective that bridges the digital divide and privacy domains. We hope that this study will also offer practical insights for policymakers and organizations seeking to address privacy inequities as a social justice issue.
Jia Wei
TAKEAWAYS
Virtual influencers can drive luxury brand attitudes via social presence. Message quality matters more for high- involvement consumers. Parasocial bonds amplify engagement and purchase intention. Realism enhances perceived authenticity in digital endorsements. AI influencers reduce reputational volatility for brands.
TAKEAWAYS
We indicate that privacy is a multilevel issue spanning society, organization, and individuals. We propose the concept of the information privacy divide across multiple levels. We suggest that access to privacy is necessary but insufficient for control over information.
Coles Research Magazine | Coles Research Symposium
29
Made with FlippingBook - Online catalogs