IaraMed: A Women’s Healthcare Chatbot for Portuguese Speakers
Fernanda B. Färber, Julia Soares Dollis, Pedro S. F. B. Ribeiro, Iago A. Brito, Rafael T. Sousa, Arlindo Rodrigues Galvão Filho · 2025
The advent of large language models has significantly advanced natural language processing, revolutionizing numerous applications within the healthcare domain. Despite these advances, most existing research remains predominantly centered around English, resulting in notable disparities in medical AI accessibility for non-English speaking communities. To bridge this gap, we introduce a specialized Portuguese-language chatbot tailored explicitly to women’s healthcare needs, addressing critical shortages in linguistic resources and culturally relevant data. Leveraging retrieval-augmented generation, our chatbot integrates accurate, evidence-based information directly into generated responses. Evaluations demonstrate that our approach markedly enhances reliability, highlighting the potential for tailored AI applications to significantly improve healthcare accessibility and outcomes for Portuguese-speaking women.