In collaboration with Dr. Sharifa Sultana, Adya Daruka, and Yang Hong

https://arxiv.org/abs/2602.00243

RQ: To what extent do LLMs generate questions about reproductive wellbeing topics in a socially sensitive way, while still being educational for it's audience?

This project aims to address the stigma surrounding reproductive wellbeing through an innovative Android application that integrates artificial intelligence for interactive education. Developed using Flutter and Dart, the app leverages OpenAI’s large language models to analyze user-submitted articles on topics such as menstrual health, contraception, and sexual health, generating multiple-choice questions (MCQs) that promote understanding while maintaining cultural sensitivity. Users can view both the full article and an AI-generated summary, answer personalized MCQs, and receive immediate feedback to reinforce learning. A built-in reporting feature allows users to flag misinformation or offensive language, with flagged content reviewed by healthcare experts or rephrased by the AI to ensure accuracy and respect.

Following IRB approval, we conducted interviews and focus groups to evaluate the app’s usability, educational effectiveness, and impact on stigma reduction, refining the design based on user feedback. With a fully functional minimum viable product, our research demonstrates that combining natural language processing with accessible mobile technology can enhance learning, foster open dialogue, and reduce stigma around reproductive wellbeing in a supportive and inclusive environment.

From Fall 2025 for UIUC Trick or Research: