The gendered lens of AI: examining news imagery across digital spaces

Abstract

This study investigates gender representation in artificial intelligence (AI)-related images across various digital spaces to understand potential biases and visual narratives in the AI domain. We analyzed a dataset of 28,199 images from news media, technology news websites, social media, knowledge-sharing platforms, and other digital spaces. Our findings revealed the prevalence of male faces and the consistent underrepresentation of women across digital spaces. We also found distinct patterns in the visual framing of men and women, with women often portrayed as being disempowered and adhering to traditional gender stereotypes. Furthermore, our cluster analysis demonstrated consistent patterns of gender representation across various visual themes, reinforcing the pervasive nature of gender biases in AI news imagery. In conclusion, our study underscores the need for conscious efforts to promote a more balanced and inclusive portrayal of gender in AI news reporting, calling for a broad societal effort toward advancing gender equality and diversity.