A multi-language Diagnosis Medical Chatbot Using Artificial Intelligence in different types of diseases

Abstract

With the growing integration of artificial intelligence in the healthcare sector, there is an increasing need for intelligent tools that support early diagnosis and enhance healthcare accessibility. This study presents the development of a multilingual medical chatbot (Arabic, French, and English) designed to act as an interactive assistant that receives simple inputs such as symptoms, age, and gender, and responds with a possible diagnosis, a brief description of the condition, and a trusted medical reference link. To achieve this, several state-of-the-art language models were applied. For Arabic, the chatbot leverages BERT Base Arabic, MARBERT, T5-Small, and GigaBERT. For English, it incorporates BioBERT, BioClinicalBERT, BlueBERTPubMed , and GigaBERT. As French lacked domain-specific models, the chatbot uses translated outputs of the best-performing English model (GigaBERT) to provide reliable responses in French. The chatbot serves as a decision-support tool for physicians by enabling quick initial assessments, while also empowering patients with accessible health insights—especially valuable in early stages before visiting a doctor or in resource-limited or multilingual environments. This project addresses the demand for digital solutions that are easy to use, linguistically inclusive, and medically relevant. It highlights the potential of AI to act as a supportive partner in delivering more efficient, accurate, and accessible healthcare for both professionals and the public.