Purpose The study aims to investigate how artificial intelligence (AI) image generators portray gender across different professional roles by examining their outputs. Moreover, the study aims to detect any patterns or discrepancies in how male and female genders are visually represented within the outputs of these generators. Design/methodology/approach The study systematically compares the gender representations of images produced by five AI image generators, namely, Ideogram, Leonardo AI, Meta AI, Stable Diffusion and Tensor Art. The analysis focuses on the depiction of images of ten distinct professions, namely, doctor, engineer, journalist, judge, librarian, nurse, pilot, professor, scientist and teacher. The generated images were analyzed using content analysis with manual annotation, where each image was labeled as male or female. Findings The findings discovered that AI-generated images perpetuate gender stereotypes across various professions. For instance, roles such as nurses and librarians are predominantly depicted by women, while men are more commonly shown in high-status or technical positions like doctors, professors, engineers, pilots and scientists. This pattern reinforces the stereotype that caregiving and supportive roles are inherently feminine, while high-status or technical roles are suited to men. Such portrayals not only reinforce societal stereotypes about gender roles but also have the potential to shape public perceptions and career aspirations, possibly deterring individuals from pursuing non-traditional careers. Originality/value To the best of the authors’ knowledge, this study is the first of its kind to make a detailed comparison of lesser-studied AI Image generators. The study is important, timely and highly relevant to global debates about AI fairness, representational ethics and digital knowledge systems. By revealing how AI-generated images portray gender within various professional roles, it highlights potential biases and stereotypes embedded in current image-generation systems. The findings can inform developers and researchers seeking to improve the fairness and accuracy of AI models, while also contributing to the development of ethical guidelines for responsible AI design. In addition, the study provides valuable insights for policymakers and organizations working to establish standards that promote equitable representation in digital media.
Mir et al. (Thu,) studied this question.