A virtual influencer explaining a meta-analysis to two million followers is no longer a thought experiment. As synthetic personas move from fashion campaigns into science communication, they inherit a question the influencer industry never fully resolved: what, exactly, are audiences trusting when they trust a face that does not exist?
Source credibility research gives us a starting vocabulary — expertise, trustworthiness, attractiveness — but virtual academics scramble the categories. The persona has no credentials, yet the institution animating it may have excellent ones. Disclosure of the persona’s synthetic nature helps audiences calibrate, but our early data suggest the timing of that disclosure matters as much as its presence.
Audiences do not extend trust to the avatar. They extend it through the avatar, to whatever they believe stands behind it.
This has practical stakes for universities experimenting with AI-fronted outreach. If the animating institution is visible and accountable, synthetic presenters can extend scientific authority to audiences that traditional formats never reach. If it is obscured, the same technology becomes an authority-laundering machine.
The research agenda, then, is not whether virtual academics are credible, but under what disclosure architectures credibility transfers legitimately — and under which ones it is merely simulated.