This study evaluated ChatGPT-4's ability to generate personalized, evidence-based nutrition recommendations in English, Russian, and Kazakh for Central Asian populations, finding moderate performance in English and Russian but very limited effectiveness in Kazakh due to language limitations. The results underscore the need for customized AI models tailored to underrepresented languages and local cultures to ensure equitable and relevant health solutions for diverse communities.
European Journal of Obstetrics and Gynecology and Reproductive Biology, Volume 306, March 2025