Bridging Language Barriers in Education: A Literature Review on Multilingual and Code-Mixed Chatbot Technologies

T. Harika , Kaniz-E-Fatima , Sofia Samreen

Multilingual chatbots, Code-mixed, NLP, language-agnostic, AI, retrieval-augmented generator (RAG), Cross-lingual communication, Conversational AI for education, Human-in-the-loop support systems, Automation of Student assistance.

AI technological advancements combined with NLP advancements drove intelligent automated student support to replace digital higher education help services. Educational institutions use question generation tools extensively for repetitive information requests in areas such as placements and exams and educational funding and academic schedule information and student accommodation policies and student support services. Heavy academic situation demand creates delays which unites email helpdesk requests with offline help desks and phone helpline requests unable to support every student adequately. Most present-day conversational systems remain English-centric which means limited language capability for their users. Current chatbot technologies create incorrect responses in India's multilingual student communication because students speak code-mixed language but also use informal terms and transliterated writing styles. Throughout this literature review we study newer advancements toward multilingual and code-mixed conversational AI systems and present important methodologies like transformer-based language modelling with cross-lingual semantic representation for token-level language identification and Retrieval-Augmented Generation (RAG). The review presents the advantages and limitations of current multilingual educational chatbots and examines persistent research challenges including translation noise together with dialect handling and language switching inconsistencies and context retention issues and absence of document-backed reasoning for institutional reliability. The study finds that academic support systems in this field require combination of language processing components with human review modules and contextual memory-based storage for reliability and universal support. Research suggests the essential need for building language-neutral chatbots for students which support continuous conversation across regional languages and incorporation of code-mixed inputs between users. Under its active operation the system effectively expands educational access alongside streamlining administrative work for universities and delivers reliable data to higher-education institutions nationwide

Volume 10 Issue 12, December-2025

Pages : a31-a39