Synthetic diversity: Racial depictions influence consumer responses to AI ‐generated models in advertising

As brands seek to increase racial representation in their advertising, many are turning to AI‐generated images of humans (i.e., synthetic models) as a scalable solution. This research demonstrates that consumer reactions to synthetic models are asymmetric, depending on the racial identity being depicted. Across five studies, we demonstrate that disclosing the AI‐generated nature of models lowers firm evaluations when the featured models are from racially underrepresented groups (e.g., Black, Hispanic, and Asian models), but not when they are White. This effect is driven by reduced perceptions of the authenticity of the firm's inclusion efforts, which in turn elevate inferences of manipulative intent. Importantly, these negative effects can be mitigated when consumers are informed that the firm is owned by members of the depicted racial community, suggesting that credible signals of authentic community ties can neutralize the evaluative penalty of AI disclosure. Our findings extend research on racialized depictions in advertising by identifying synthetic depiction as a novel threat to the evaluative benefits of minority representation, and by theorizing why this threat operates asymmetrically by race. More broadly, our results highlight that technological innovations in advertising cannot be divorced from the social and historical contexts that shape their meaning.