Discussion about the feasibility of the employment of natural language processing tools in the identification of binding decisions in judicial processes

Authors

  • Daniel Henrique Arruda Boeing
  • Luccas Fernandes de Quadros
  • Tiago Ribeiro Alves de Melo
  • Renata Matos

Abstract

This paper discusses the use of Natural Language Processing (NLP) tools on the identification of binding precedents themes in legal documents. Therefore, there are legal and technical explanations. From the legal standpoint, the paper examines the: (a)  Binding Precedents System instituted by the Brazilian Code of Civil Procedure from 2015; (b) the new legal mindset required to its adoption; and (c) how it aims to provide more celerity, isonomy and legal security to judgements. From the technical standpoint, NLP tools are presented in order to analyse the feasibility of their implementation on the identification of the convergence between the precedent and the document as a support to expert agents’ decision making. Finally, dynamics and results from LEIA Precedents are presented, a real-case project through which this strategy was successfully carried out.

Author Biographies

Daniel Henrique Arruda Boeing

Graduado em Direito pela Universidade Federal de Santa Catarina. Analista de Dados no Departamento de Inovação da Softplan. E-mail: danielboeing96@gmail.com.

Luccas Fernandes de Quadros

Graduado em Ciência da Computação pela Universidade Estadual Paulista Júlio de Mesquita Filho. Cientista de Dados na empresa Softplan. E-mail: luccas.qdrs@gmail.com.

Tiago Ribeiro Alves de Melo

Doutor em Economia Organizacional pela Universidade de Salamanca, Espanha. Mestre em Mestrado em Business Management pela Universidade de Liverpool, Reino Unido. Product Manager na empresa Softplan. E-mail: tiago.melo@softplan.com.br.

Renata Matos

Graduada em Direito pela Universidade do Sul de Santa Catarina. Advogada. E-mail: advrenatamatos@gmail.com.

Published

2020-11-26