AI-Powered Celebrity Substitution After Endorser Scandal: Promotion-Governance in Convergent VR-AR-AI Media Ecosystems

This study examines consumer acceptance of AI-powered celebrity surrogate substitution as a scandal management mechanism within convergent VR-AR-AI marketing ecosystems. Using PLS-SEM analysis of 221 Vietnamese consumers screened for awareness of a real-world AI substitution case, the study tests five antecedents of attitude toward the AI surrogate (ATAS) and its effect on viewing intention. Grounded in Image Repair Theory and Schema Incongruity Theory, findings show that residual positive affect toward the scandalized celebrity transferred to surrogate evaluations, while stronger perceived anthropomorphism significantly reduced acceptance reflecting schema-based symbolic contamination. Moral reasoning constructs showed no significant effect. ATAS strongly predicted viewing intention and mediated key upstream paths. The study distinguishes AI surrogates from virtual influencers and digital resurrections and identifies a novel boundary condition: anthropomorphism’s effect reverses from positive to negative in post-scandal contexts. Practitioners should adopt visible AI editing, implement proactive disclosure across platform touchpoints, and pursue parallel reputation management strategies.