Mina: A Multilingual LLM-Powered Legal Assistant Agent for Empowering Access to Justice in Bangladesh

Azmine Toushik Wasi, Wahid Faisal, Mst Rafia Islam, Md Rizwan Parvez

Keywords: multilingual / low resource; applications; legal NLP, NLP for social good; less-resourced languages; LLM/AI agents; valuation and metrics; task-oriented; human-in-the-loop; bias/toxicity; factuality; retrieval; knowledge augmented;

Abstract: Bangladesh’s low-income population faces major barriers to legal assistance due to complex language, procedural opacity, and high costs, while existing AI legal systems lack Bengali and jurisdiction-specific support. To address this gap, we introduce \textsc{Mina}, a multilingual LLM-based legal assistant for Bangladesh that integrates multilingual embeddings with a RAG-based chain-of-tools framework for legal retrieval, reasoning, translation, citation insertion, and document drafting. Evaluated by law faculty experts across all stages of the 2022 and 2023 Bangladesh Bar Council examinations, \textsc{Mina} achieved 75--80\% scores in MCQ, written, and simulated viva evaluations, matching or surpassing average human performance in legal reasoning and contextual understanding. Operating at only 0.1--0.6\% of the cost of traditional legal services, \textsc{Mina} demonstrates the feasibility of scalable, low-cost multilingual legal assistance for underserved populations in low-resource settings.

Track: Track 2: ML Research by Muslim Authors

Email Sharing: We authorize the sharing of all author emails with Program Chairs.

Data Release: We authorize the release of our submission and author names to the public in the event of acceptance.

Non Archival Confirmation: I understand that submissions to MusIML are non-archival and can be submitted to other venues.

Submission Number: 29