Key points are not available for this paper at this time.
Generative AI is reshaping advertising. Virtual influencers are a salient example, functioning as AI-generated advertising faces across digital platforms. This systematic review synthesized 159 peer-reviewed studies using the Theory – Context – Characteristics – Methodology (TCCM) framework. The literature relied heavily on human-centered persuasion theories, mainly source credibility and parasocial interaction. Virtual influencer effects were largely indirect. They operated through cognitive, emotional, and relational mechanisms and depended on contextual and individual moderators. Short-term quantitative designs dominated the field. The review proposes an integrative framework. It maps a three-stage process from influencer cues, through mediating mechanisms, to behavioral, engagement, and authenticity outcomes, moderated by congruence, consumer, and contextual conditions. The framework also specifies what current persuasion, influencer, and human – computer interaction theories cannot explain about non-human endorsers. The review identifies three tensions: the undertheorized role of CASA theory relative to Parasocial Interaction Theory, the need to distinguish heterogeneous virtual influencer types, and the absence of frameworks addressing algorithmic agency.
Phan et al. (Wed,) studied this question.