CARE-AI: Culturally Adaptive and Responsible AI for Equitable Health Communication

Artificial Intelligence (AI) is playing a bigger role in how patients and healthcare systems communicate. While these tools can improve efficiency, they often struggle with cultural and linguistic differences. Most AI systems are trained primarily on English-language medical data, which means they can overlook local traditions, values, and community-based care practices found in places like Southern Europe or Latin America. When AI fails to understand these cultural nuances, it can lead to misunderstandings, especially for migrant or multilingual families, ultimately damaging trust and leading to poorer health outcomes - with particularly high stakes in maternal health. This project aims to make healthcare AI more inclusive by building a bilingual (English–Spanish) database using real-world maternal care communications. Using this data, we will test a new interface to see if AI can be trained to recognize and respect cultural differences. Our goal is to ensure that healthcare technology works fairly for everyone, regardless of their background or the language they speak.