Virtual influencers, algorithmically designed brand-controlled personas, enact persuasion without experiential intention, moral agency or accountability. Existing constructs have not captured act-level evaluation of such conduct. This study introduces Perceived Persuasion Legitimacy of Virtual Influencers (PPL-VI), examining associations with willingness to accept influence (WAI) and purchase intention (PI).
Study 1 developed PPL-VI through cue mapping, focus groups (N = 30) and expert interviews (N = 10). Study 2 assessed it via a post-exposure survey (N = 878) of Indian social media users, using PLS-SEM with a Type II hierarchical specification. Study 3 (N = 214) was a two-condition experiment.
PPL-VI comprises four formative dimensions: Integrity of Persuasion Enactment, Respect for Persuasive Boundaries (RPB), Ethical Treatment of Consumers and Clarity of Responsibility for Persuasion, with RPB carrying the greatest outer weight in Study 2. Value alignment partially mediates the PPL-VI–PI association, procedural fairness the PPL-VI–WAI association and perceived human-likeness modestly moderates both. Study 3 suggests PPL-VI is associated with variance in purchase intention beyond adjacent constructs under lower-legitimacy conditions.
Study 2 is cross-sectional, so causal inference requires further replication. The India-only, non-probability samples limit generalisability; Study 3 manipulated one dimension.
Brands may benefit from monitoring boundary restraint and ethical consumer treatment alongside the other dimensions; recommendations are hypothesis-generating and directional rather than prescriptive.
This study develops and provides initial evidence for a 16-item instrument for act-level legitimacy evaluation in AI-mediated persuasion and proposes an act-level PKT adaptation integrating the Computers as Social Actors paradigm.
