Pakistani student builds AI legal assistant HAQ using RAG for accessible law

A computer science student in Pakistan has developed HAQ, an AI-powered legal assistant designed to make Pakistani law accessible to the general population. Unlike standard LLMs that can confidently provide incorrect legal information, HAQ utilizes Retrieval-Augmented Generation (RAG) to ensure answers are based solely on retrieved passages from actual Pakistani legislation. The system retrieves relevant legal text before the AI generates an answer, linking directly to the source document and explicitly stating when no relevant information can be found to prevent dangerous hallucinations. The project addresses the significant challenge of accessing legal information due to language barriers, high consultation fees, and the digitization issues of legal documents in Pakistan. AI

IMPACT Could significantly improve access to legal information for underserved populations by providing reliable, source-backed answers in local languages.

RANK_REASON AI-adjacent product launch by a student.

Read on dev.to — LLM tag →

• HAQ
• LLMs
• Pakistani law
• Retrieval-Augmented Generation

• product
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →