<p>Role models influence individuals’ aspirations, values, and societal norms, shaping career choices and academic pursuits. Counter-stereotypical role models could challenge traditional norms and promote diversity and inclusion. As Artificial Intelligence (AI) increasingly shapes social perceptions, this study investigates gender bias in AI-generated definitions of role models. Using prompts in seven languages, English, Greek, Hebrew, Urdu, Somali, Kurdish, and Punjabi, we examined how AI defines female, male, and gender-unspecified role models, with a focus on linguistic and cultural variation in the outputs. Through discourse and content analysis of AI-generated descriptions, we identified persistent gender biases across all languages. However, patterns varied significantly based on cultural and linguistic context. In languages spoken in societies with lower gender equality, such as Punjabi, Urdu, Kurdish, and Somali, female role models were more often portrayed with nurturing, moral, and family-oriented traits, while male role models were associated with ambition, authority, and public achievement. By contrast, English and Greek outputs, showed more overlap in the traits attributed to male and female role models, reflecting a more Western approach to gender equality in discourse. Importantly, even in these higher-equality contexts, gender bias remained evident, especially in the default assumption of male identity in gender-unspecified prompts, particularly in Greek due to grammatical gender. These results highlight how cultural and linguistic factors influence AI-generated discourse, often reinforcing localized gender norms. Our findings underscore the need for bias mitigation strategies in AI systems to promote more inclusive and equitable representations, contributing to fairer societal perceptions and greater gender equality in AI-generated content.</p>