This person does not exist: Detailed AI disclosure and trust in synthetic visual brand posts

Brands increasingly use hyper-realistic AI-generated human figures in social media advertising, raising questions about how disclosure affects consumer trust once audiences learn the depicted person does not exist. Drawing on the Persuasion Knowledge Model and Signaling Theory, this study examines how prior AI disclosure shapes the reassessment of message trust after consumers receive external information about the visual’s synthetic origin. An online experiment (N = 606) exposed participants to a brand’s Instagram post under one of three conditions: no disclosure, generic disclosure (‘AI-generated’), or detailed disclosure (explicitly stating the depicted person does not exist), and measured trust before and after external information about the image’s synthetic nature. Results show that detailed disclosure’s effect on trust operates entirely through two opposing mechanisms rather than directly: it reduces inferences of manipulative intent, thereby protecting trust, while simultaneously increasing perceived syntheticness, thereby eroding trust. Age moderates the manipulative-intent pathway, with younger consumers showing stronger protective effects, but does not moderate syntheticness. Findings advance understanding of disclosure specificity’s dual-edged role in fully synthetic visuals and offer practical guidance for brands and policymakers navigating AI content transparency requirements.