Original research

LEGALEASE: AI-POWERED SUMMARISATION AND VISUALIZATION OF LEGAL DOCUMENTS

Pages 59-66

10.61552/JAI.2026.01.006

Siddhi Bhogale, Nidhi Nayak, Shivani Nikam, Pranathi Narsupalli, Vidya Pujari

• View Article
• PDF 823 K

Abstract Legal documents are often complex and filled with specialized jargon, making them difficult to comprehend. LegalEase is an AI-driven tool designed for effective summarization and visualization of legal texts using mind maps. It fine-tunes the T5-small model on datasets like IN-ABS and MILDSum to generate concise, contextually aware summaries. Additionally, it utilizes the LlaMA model for structured formatting, D3.js for flowchart-based visualizations, and Google Cloud API for multilingual translation, including Hindi. The system also features jargon simplification using a JSON-based dictionary to enhance readability. Experimental results show significant improvements in summarization quality, with enhanced ROUGE, Meteor, and BERTScore metrics compared to pre-trained models. By integrating visual and multilingual tools, LegalEase improves accessibility and usability across various legal domains and audiences.

Keywords: Legal document summarization, T5-small, Natural Language Processing, Abstractive summarization, Large Language Model.

Recieved: 21.02.2025. Revised: 03.05.2025. Accepted: 18.06.2025.