Despite the increasing role of artificial intelligence (AI) in generating visual content, concerns regarding the biases embedded within generative models persist. This study analyses the representation of the academic student community in images generated by the Flux.1 model. Two sets of N = 1000 images in total, covering M = 10 field of studies, were generated. This was achieved using two distinct text queries: ‘A [field of study] university student’ and ‘A university [field of study] student’, which allowed the effect of sentence formation on the final output of the model to be assessed, including any potential bias associated with it. The Grok 3 AI tool facilitated the analysis, focusing on the identification of: gender, facial features, clothing, accessories, general appearance, and interaction with the surroundings of the portrayed students. The results indicate a clear gender bias based on stereotypical field perception, manifested by the underrepresentation of women in the usually more analytically oriented fields. Furthermore, the model often reinforced stereotypical visual characteristics, such as long hair for women or facial hair for men. The findings emphasise the necessity for greater diversity in training data and responsible AI development to mitigate biases in academic and professional representation.