Aspectos éticos e regulatórios para a implantação de soluções baseadas em inteligência artificial no poder judiciário brasileiro

Abstract

Artificial intelligence solutions are increasingly taking a prominent place in people's daily lives. Today, they are integrated into various fields of knowledge, including the legal sphere, with a significant presence in the judiciary. With the widespread use of AI, other important concerns arise, such as ethics, responsibility, and transparency. Therefore, the objective of this dissertation is to assess the ethical and regulatory aspects related to the implementation of AI-based solutions in the Brazilian judiciary. To accomplish this, an analysis was conducted using data obtained through a questionnaire administered by the Federal Court of Accounts in process TC 006.662/2021-8, as well as a systematic literature review using the PRISMA method. Issues related to justice, responsibility, and transparency in the development of artificial intelligence in 45 federal judiciary courts were investigated, considering the framework known as "FAT" (Fairness, Accountability, and Transparency). The analysis also addressed aspects related to technical and structural support within organizations for AI implementation. Initially, a literature review was conducted on artificial intelligence and the "FAT" framework, which serves as a minimum ethical model for the implementation of AI-based solutions. Subsequently, the discussion focused on Resolution No. 332/2020 of the National Council of Justice, which deals with ethics, transparency, and governance in the production and use of AI in the judiciary. This resolution establishes guidelines for the development and use of AI and the graphs with data obtained by the Federal Court of Accounts were analyzed. The systematic literature review aimed to identify the main challenges in the ethical implementation of AI systems and the primary risks associated with the use of AI that does not adhere to ethical parameters. The responses provided by the courts indicate a lack of adherence to the FAT framework, as the criteria related to justice, responsibility, and transparency were not observed by the questionnaire participants. This also results in non-compliance with relevant aspects of Resolution No. 332/2020-CNJ, which is aligned with the FAT framework. Furthermore, through the systematic literature review, an urgent need for AI regulation and the establishment of clear standards for its implementation was identified. This is crucial to make the definition and application of ethical principles accessible and transparent. Therefore, it is concluded that the use of AI without appropriate ethical parameters can pose significant risks to individuals, including violations of fundamental rights and the perpetuation of inequalities and discrimination.