Bridging Languages in Healthcare: A Comprehensive Review of Multilingual and Code-Switched Chatbot Interactions

The Growth of societies with multilingual populations and the expanded accessibility to the healthcare service digitally has led to an increasing demand. For intelligent conversational agents that can operate and manage interactions with complex languages. Especially in countries like India, there exists a common process of code-switching between English and the regional or local languages, where the monolingual system falls conventionally. The paper explores the healthcare chatbots in the current state of research and the developments in the response and detection of queries, which involve multilingual and code-switching. This review aims to analyze the detection and response in the existing approaches to multilingual and healthcare queries. Also, the objective is to identify strengths, limitations, and research gaps in the current methodology. We highlight the challenges produced by the variation in the linguistic, languages with low resources, and the ambiguity of the context. Our analysis highlights that the performance of transformer-based multilingual models remains limited in low-resource and highly code-mixed settings, whereas the robustness of the models is improved. The paper analyses the models that exist for the identification of languages, detection of code-switching at the token-level, multilingual embedding, and systems based on transformers. We further discuss evaluation benchmarks, moral considerations, with ethical factors to direct future works in the equity of healthcare enhancement through AI with inclusive language.