Aesthetically Ideal Breasts Created With Artificial Intelligence: Validating the Literature, Racial Differences, and Deep Fakes.

Abstract

A female's breasts are integrally tied to her identity and sense of femininity. Despite extensive study of breast aesthetics, there is no discrete formula for the "ideal breast" to guide the aesthetic surgeon. Racial and cultural differences heavily influence preferences in breast morphology. Artificial intelligence (AI) is ubiquitous in modern culture and may aid in further understanding ideal breast aesthetics. This study analyzed AI-generated images of aesthetically ideal breasts, evaluated for morphologic differences based on race, and compared findings to the literature. An openly accessible AI image-generator platform was used to generate images of aesthetically ideal Caucasian, African American, and Asian breasts in 3-quarter profile and frontal views using simple text prompts. Breast measurements were obtained and compared between each racial cohort and to that of previously described ideal breast parameters. Twenty-five images were analyzed per racial cohort, per pose (150 total). Caucasian breasts were observed to fit nicely into previously described ideal breast templates. However, upper-to-lower pole ratios, nipple angles, upper pole slope contours, nipple-areolar complex positions, and areolar size were observed to have statistically significant differences between racial cohorts. Defining the aesthetically ideal breast remains a complex and multifaceted challenge, requiring consideration of racial and cultural differences. The AI-generated breasts in this study were found to have significant differences between racial groups, support several previously described breast ideals, and provide insight into current and future ethical issues related to AI in aesthetic surgery.