Coded femininity : AI, beauty standards, and the commodification of the female body in fashion

Abstract

The paper examines how Artificial Intelligence encodes the female body, transforming a historically fluid, culturally shaped concept of beauty into a narrow, digitally “measurable” ideal. Generative AI, trained on biased fashion industry datasets, reproduces the Western-centric model of young, thin, white and symmetrical femininity, eliminating diversity and establishing a normative aesthetic. Through virtual try-ons, filters, AI avatars and virtual influencers, the body is transformed into a digital asset and social currency, while users are pushed into constant self-surveillance, comparison and “optimisation”, with measurable consequences on self-image, mental health and consumer practices. The biases hidden behind algorithms, such as thinness, skin colouration and hyper-sexualisation, do not simply mirror existing social exclusions. On the contrary, they reinforce them, offering greater visibility and imitability, while deepfakes add even more complexity. Thus, artificial intelligence is transformed into a tool that disciplines and controls the image of women. It is imperative to have transparency, use multi-collective datasets, conduct ethical programming and implement regulatory interventions. Only in this way can we ensure that the use of AI in the fashion and beauty sector does not continue to reproduce aesthetic violence and commercialisation.