Using AI in the Public Sphere
With a current backlog of 78 million lawsuits, the Brazilian judicial system operates with substantial challenges in caseflow management and a lack of resources to meet this demand. In order to address these issues, the Brazilian National Council of Justice (CNJ) has enabled the 92 courts it administratively oversees to develop their own artificial intelligence models, resulting in a seemingly uncoordinated algorithmic universe in the judicial system.
In spring 2020, a SIPA Capstone team partnered with the Institute for Technology and Society (ITS)—a nonprofit research institute based in Rio de Janeiro—to design a collaborative governance structure to strategically integrate all AI initiatives in the Brazilian judiciary.
The team—Katie Brehm MIA ’20, Momori Hirabayashi MPA ’20, Clara Langevin MPA-DP ’20, Bernardo Rivera Mu.ozcano MPA ’20, Katsumi Sekizawa MPA ’20, and Jiayi Zhu MPA ’20—interviewed a sample of experts in the field, conducted a literature overview of AI and electronic processing systems in the Brazilian judiciary, developed a comparative analysis on both national and international case studies, and, ultimately, compiled a list of strategies and recommendations for the CNJ.
“The quality of the report is impressive,” says Ronaldo Lemos, director of ITS Rio and a SIPA adjunct senior research scholar. “Their research skills are outstanding, and the level of detail with which they analyzed the case studies imagazine article s s evidence of their analytical rigor. The potential impact of this work is enormous, since it may be used as a guidance to reshape the entire Brazilian judiciary system.”
The report’s recommendations include adoption of an agenda regarding the use of AI tools within the Brazilian judiciary, creation of an AI and assessment tool, increasing the judicial system’s transparency through the use of open-source software, improving AI interoperability among courts’ systems, increasing court collaboration, and building and disseminating a framework for the provision or procurement of AI tools from the private sector by courts.
“The Brazilian judicial system is quite complex, and the project team managed to reach a high level of understanding of its reality, its configuration, and its major problems and challenges faced nowadays,” adds Luiz Antonio Mendes Garcia, chief information officer at the CNJ.
The workshop, The Future of AI in the Brazilian Judicial System: AI Mapping, Integration, and Governance, was advised by MPA-DP assistant director and adjunct associate professor André Corrêa d'Almeida.
“As a Brazilian, it was a great opportunity to work directly with both the judicial branch and a renowned research institute,” Langevin says. “I am proud of the work that we did and hope that it will continue to inform the use of AI in the public sphere, especially at this moment where the country is seriously considering a national-level AI strategy.”
This story appears in the most recent issue of SIPA Magazine, published in October 2020.