Breaking Gender Stereotypes? Analyzing Gender Representations in AI-generated Fashion Campaigns

Abstract: Advertising has a long history of perpetuating gender stereotypes. In earlier forms of advertising, women were predominantly portrayed in the roles of housewives and caretakers, and men in a professional environment and as breadwinners. Similarly, the fashion industry is no exception to making use of gendered stereotypes, with fashion campaigns and advertising often portraying models as exceptionally beautiful and disturbingly thin, which hardly corresponds to reality. However, the advent of artificial intelligence (AI) allows unprecedented opportunities for the creation of fashion campaigns, potentially enabling fashion brands to create more diverse and inclusive campaigns that ultimately break away from gender stereotypes. This study empirically assesses the state of current gender representation in the fashion industry by examining promotional campaigns created with AI. By applying Goffman-inspired content coding of gender representations, the study shows that AI-generated fashion campaigns challenge stereotypes by depicting virtual AI models fully clothed, with confident, independent postures, and with various ethnicities, yet they still reinforce gender stereotypes in their more subtle forms.