Kim et al. (2023) — The next hype in social media advertising: Examining virtual influencers’ brand endorsement effectiveness

Kim, E.-J., Kim, D., E, Z. & Shoenberger, H. (2023). The next hype in social media advertising: Examining virtual influencers’ brand endorsement effectiveness. Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2023.1089051


What this source contributes

Tests the effectiveness of virtual influencer endorsements by comparing human-like virtual influencers (HVIs) against anime-like virtual influencers (AVIs). The theoretical frame is the CASA paradigmComputers Are Social Actors — which holds that people automatically apply social rules to computer-generated agents that appear sufficiently human-like. The more human the appearance, the more the agent triggers automatic social trust responses, increasing perceived social presence and message credibility.

The key finding is that HVI endorsements produce greater perceived message credibility than AVI endorsements, but only when sponsorship is not disclosed. Once AI origin is made explicit, the trust advantage disappears entirely.

Analytical function in Virtual Influencer

Provides the CASA paradigm as the theoretical explanation for why anthropomorphic design functions as trust engineering. The disclosure finding anchors the Friction section: the advantage exists precisely because disclosure is suppressed, which makes its burial in fine print a mechanism rather than an oversight.

  • Virtual Influencer — the CASA paradigm as the explanation for why human-like design produces trust