With 47 million pending cases and only 21 judges per million population, India is facing a serious crisis of legal access, which is disproportionately affecting the rural population. This Paper proposes an AI- powered multilingual legal aid platform called Nyay Sahayata, which generates court-compliant documents (FIRs, RTIs, affidavits) in over 22 Indian languages. Our system uses a novel Retrieval Augmented Generation (RAG) framework along with IndicBERT and AI4India's IndicTrans2 for accurate multilingual legal processing. The estimated performance analysis on the proposed framework indicates a document acceptance efficiency of 95.2% from legal authorities, reduced filing time from 15 days to an average of 12.7 minutes, and demonstrated 92.8% accuracy in multilingual legal vocabulary processing. Voice input functionality is designed to help users with low literacy independently prepare legally valid documents. Our findings suggest that AI-powered multilingual legal assistance could significantly improve access to justice for disadvantaged populations. The RAG framework ensures legal accuracy while removing barriers to voice input literacy. This paper proposes a comprehensive framework for voice-enabled multilingual legal AI tailored to developing regions, with the potential to improve global access to justice.
Pardey et al. (Fri,) studied this question.