Realistic Portrayals of Untrue Information: The Effects of Deepfaked Ads and Different Types of Disclosures
通过三项实验，研究了深度伪造广告如何影响消费者，以及不同披露方式（简单披露与扩展披露）在降低其误导性上的效果，发现扩展披露能降低感知真实性但无法降低呈现真实感。
A deepfake is a realistic portrayal of untrue information (i.e., it realistically depicts individuals who were not originally present in a given footage). A deepfaked ad has the potential to mislead consumers because it shows people saying or doing things that they have never said or done, which can harmfully influence consumer decisions. Despite the increasing prevalence of deepfakes in various contexts, there has been no research on their implications in advertising and how to effectively protect consumers through disclosures. Therefore, we conducted three studies. Study 1 shows that consumers recognize deepfakes as synthetically generated ads only if there is a clear indication or disclaimer regarding the synthetic nature of deepfakes. Study 1 also demonstrates that simple disclosures (informing about the alteration of the original message) do not reduce the perceived portrayal realism of deepfakes. Study 2 examines new extended disclosures for deepfakes (informing about the consequences of alteration for an ad and a consumer). The study shows that these disclosures are successful in reducing the perceived factuality, but not the portrayal realism, of deepfakes. Study 3 replicates this effect and tests a boundary condition (i.e., prior expectations about advertising factuality).