CoLABOR uses large language models (LLMs) to analyze the case law of the Supreme Administrative Court (STA)
CoLABOR supports the analysis of the case law of the Supreme Administrative Court (STA) by automatically extracting structured information from court decisions using artificial intelligence models.
This initiative, which is part of the project “Tax Litigation in Portugal: A Legal and Empirical Study” carried out by CIDEEFF, will make it possible to organize and analyze large volumes of case law, contributing to a more comprehensive understanding of tax litigation in Portugal.
The classification of court decisions is based on a set of predefined parameters and follows an iterative methodology designed to ensure the quality, consistency, and reliability of the information produced. The process useslarge language models, whose results are validated by comparing them with a set of court decisions previously classified by experts.
The main outcome of the project—which will last four months—will be a structured database that will enable large-scale analysis of the patterns, trends, and characteristics of the STA’s decisions, serving as a highly valuable resource for researchers, legal professionals, and public policymakers.
By applying artificial intelligence technologies to legal research, the project aims to contribute to new ways of analyzing, organizing, and producing knowledge in the field of law, demonstrating the potential of these technologies to support scientific research.
As part of this collaboration, CoLABOR will support the analysis of the case law of the Supreme Administrative Court (STA) by automatically extracting structured information from court decisions using artificial intelligence models. This approach will make it possible to systematize and analyze large volumes of case law, contributing to a more comprehensive understanding of tax litigation in Portugal.
The classification of court decisions is based on a set of predefined parameters and follows an iterative methodology designed to ensure the quality, consistency, and reliability of the information produced. The process uses large language models (LLMs), whose results are validated by comparing them with a set of court decisions previously classified by experts.
The main outcome of the project—which will last four months—will be a structured database that will enable large-scale analysis of the patterns, trends, and characteristics of the STA’s decisions, serving as a highly valuable resource for researchers, legal professionals, and public policymakers.
By applying artificial intelligence technologies to legal research, the project aims to contribute to new ways of analyzing, organizing, and producing knowledge in the field of law, demonstrating the potential of these technologies to support scientific research.