Huqooqi: AI Legal Consultation App for Private Sector Workers

Title of the Project

Students Details

202210788 Maryam Dhafer Jumaah

202110646 Sarah Ali Alajmi

202120184 Fatima Waad Altaie

202120002 Sarah Mohamad Hassan Issa

Abstract

This report is a design, development, and evaluation of a bilingual UAE Labor Law Assistant, which is an application that runs on Retrieval-Augmented Generation (RAG). The project aims to deliver the correct answers to employment-related questions in Arabic and English based on the artificial intelligence and with legal foundation. The system combines the OpenAI GPT language model to build natural language with semantic similarity search using Qdrant vector database, as all the responses generated are also trackable to the official legal documents.

The frontend is developed using React and the Tailwind CSS, which is backed by the Axios, used to make API calls and styled to meet the needs of responsive user interaction. The fastapi, mongoDB to store structured data, and qdrant to search vectors are used to develop the backend. A well-organized development process, based on Agile principles, made it possible to continuously add new features and continuously test them.

Testing stage included unit, integration and end-to-end testing to confirm the performance of the backend, responsiveness of the front end and consistency of the user experience between devices and browsers. The special attention was paid to security, performance optimization, and multilingual support. The system had a high rate of response accuracy of 94-96 percent as well as a hallucination rate of less than 2 percent. Reliability requirements of legal AI systems led to the implementation of error handling, traceability and legal citation mechanisms.

The present project proves the validity of incorporating AI into legal advisory devices and provides a scalable and user-friendly method of legal awareness in the UAE labor market.