On the realness of people who do not exist: The social processing of artificial faces

Today more than ever, we are asked to evaluate the realness, truthfulness and trustworthiness of our social world. Here, we focus on how people evaluate realistic-looking faces of non-existing people generated by generative adversarial networks (GANs). GANs are increasingly used in marketing, journalism, social media, and political propaganda. In three studies, we investigated if and how participants can distinguish between GAN and REAL faces and the social consequences of their exposure to artificial faces. GAN faces were more likely to be perceived as real than REAL faces, a pattern partly explained by intrinsic stimulus characteristics. Moreover, participants' realness judgments influenced their behavior because they displayed increased social conformity toward faces perceived as real, independently of their actual realness. Lastly, knowledge about the presence of GAN faces eroded social trust. Our findings point to potentially far-reaching consequences for the pervasive use of GAN faces in a culture powered by images at unprecedented levels.