Aplicação de Large Language Models na Análise e Síntese de Documentos Jurídicos: Uma Revisão de Literatura

TL;DR

Este estudo aborda o estado da arte da engenharia de prompts em LLMs aplicados à análise de documentos jurídicos em Português. Emprega uma revisão sistemática de literatura (2020–2024) e consolida evidências sobre modelos usados (GPT-4, Llama 2, Legal-Pegasus, Sabiá, etc.) e técnicas (Zero-shot, Few-shot, Chain-of-Thought). Mostra que, embora promova desempenho superior em várias tarefas, persistem alucinações e vieses, exigindo melhorias em prompts, ajuste fino e dados especializados. Contribui para orientar pesquisadores e profissionais sobre estratégias de prompt, avaliação e adoção responsável de LLMs no direito, com ênfase na confiabilidade e conformidade legal.

Abstract

Large Language Models (LLMs) have been increasingly used to optimize the analysis and synthesis of legal documents, enabling the automation of tasks such as summarization, classification, and retrieval of legal information. This study aims to conduct a systematic literature review to identify the state of the art in prompt engineering applied to LLMs in the legal context. The results indicate that models such as GPT-4, BERT, Llama 2, and Legal-Pegasus are widely employed in the legal field, and techniques such as Few-shot Learning, Zero-shot Learning, and Chain-of-Thought prompting have proven effective in improving the interpretation of legal texts. However, challenges such as biases in models and hallucinations still hinder their large-scale implementation. It is concluded that, despite the great potential of LLMs for the legal field, there is a need to improve prompt engineering strategies to ensure greater accuracy and reliability in the generated results.