Evidence of systematic differences between the faces of social media creators and the general population

The Internet both substantially expands and potentially biases our “visual diet” concerning faces. Using an evolutionary psychology framework, we compared the facial characteristics of popular content creators on online social media with control faces from the same populations. We analysed facial configurations with landmark-based geometric morphometrics and Bayesian models, conducting the first such comparison on a large sample. Across all four samples (Colombia: N = 291, 153 creators; Czechia: N = 360, 165 creators; India: N = 201, 63 creators; Turkey: N = 407, 145 creators), creators showed similar traits, including pronounced sexual dimorphism and evenly distributed features. The analysis also consistently showed significant differences between the creators and controls. We then extended the dataset with faces generated by a text-to-image model to investigate Internet typicality, for which we found only limited support. These results indicate systematic differences between the faces of popular creators and those of the general population.