Fundação Getulio Vargas’ Center for Judicial Innovation, Administration and Research (CIAPJ) has just published the third phase of its study on artificial intelligence in the Brazilian courts, which analyzes artificial intelligence tools in courts in Brazil’s Midwest region, based on the ethics, transparency and governance criteria established in National Justice Council Resolution 332.

The center, which is coordinated by FGV professor and Justice Luis Felipe Salomão, has been monitoring artificial intelligence projects in the judiciary since 2019. In the latest phase of this research project, the scientific coordinator was Professor Dierle Nunes and the collaborators were FGV researchers Fernanda Bragança, José Leovigildo Coelho, Juliana Loss and Renata Braga.

The study explored the flow and functioning of the National Justice Council’s Justice 4.0 Program and its Sinapses, Codex and DataJud platforms, the Federal Supreme Court’s Rafa and Victor tools, the Superior Court of Appeals’ Athos system, the Superior Labor Court’s Bem-te-vi platform, the Regional Federal Court for the First Region’s ALEI tool, and the Federal District and Territories Court of Appeals’ Amon, Toth, Saref, Artiu and Horus systems.

The researchers identified that some of the AI systems adopted by the courts need to be improved in certain aspects, such as the publicity of documentation and control and security instruments, the presentation of authoring mechanisms and the certification of best practices. It is also necessary to inform external users about the use of AI and to provide explanations for decisions that can be understood by humans.

To read the full report, click here.