This study examines how GPT-generated responses influence gender-related perceptions across countries including Iran, UAE, Bangladesh, Indonesia, and Qatar. Given the sensitivity of gender issues within these societies and the growing use of AI in shaping public opinion, participants engaged in a two-phase evaluation process using a 5-point Likert scale. Initially, participants rated their agreement with statements on gender issues, followed by exposure to GPT’s responses to the same statements. Participants then rated their agreement with GPT’s perspectives, allowing for an indirect measurement of AI’s influence on their perceptions. While some responses remained stable, others, particularly those related to societal roles and gender equality, exhibited significant shifts. Using T-tests and a Dirichlet distribution, selected to capture both mean differences and distribution-level changes, the analysis revealed that GPT’s responses significantly affected perceptions, notably on topics such as legal rights and economic independence, suggesting a progressive impact. However, the influence was not uniform across genders: female participants generally reported positive shifts, whereas male participants frequently experienced neutral or negative shifts. These findings underscore the varied impact of GPT on gender perceptions and highlight the necessity for ongoing research to mitigate AI biases and effectively engage diverse groups in discussions on sensitive social issues. Insights from this study may contribute to informed policy-making and responsible AI deployment, promoting gender equity and balanced discourse.