Entre algoritmos e direitos fundamentais : a necessária articulação entre o Marco Legal da Inteligência Artificial e as políticas públicas de diversidade no enfrentamento da discriminação algorítmica de gênero no Brasil.

Abstract

The doctoral research focuses on the Legal Framework of Artificial Intelligence in Brazil and its indispensable role in combating gender-based algorithmic discrimination. Bill No. 2,338/2023 has been taken as the proposed AI Legal Framework and will be the object of analysis in this thesis, since it is at an advanced stage of legislative proceedings, rendering other related bills ineffective and leading to their archiving.The scope of this analysis is algorithmic discrimination specifically directed against women, which must be specified given that gender is a category encompassing diverse experiences and bodies. Based on this scenario, the central research question is: Can the legal regulation of Artificial Intelligence, through a Legal Framework, adequately address gender-based algorithmic discrimination, including its prevention, while respecting Fundamental Rights? The general objective is precisely to analyze this problem. The initial hypothesis is a conditional affirmative, in the sense that regulation will only be able to address this challenge if accompanied by Public Policies on diversity. The specific objectives are as follows: to conceptualize gender through feminist theory, highlighting the mechanisms through which women are socialized and the overlapping intersectional markers of oppression that intersect with the category of gender; to present gender equality as a Fundamental Right, addressing material and substantive equality, as well as equity, while examining the main international conventions on gender ratified by Brazil, together with SDG 5 of the 2030 Agenda; to develop the necessary conceptualizations for understanding Artificial Intelligence and its learning processes within the selected scope, including an explanation of how gender-based algorithmic discrimination is produced or reproduced in automated decision-making; to list the bills under consideration in the Brazilian Congress aimed at regulating AI, with special emphasis on Bill No. 2,338 of 2023 and the path leading to its formulation; and to propose diversity-oriented Public Policies that should accompany the AI Legal Framework in Brazil in order to adequately confront gender based algorithmic discrimination, including its prevention. The investigation into the existence or absence of gender-based algorithmic discrimination will be carried out through a search process using the image-generation system based on Artificial Intelligence known as “Dall-E 3.” The search terms will be constructed on the basis of linguistically neutral conceptual categories defined within the chosen theoretical framework, in order to encompass expressions that, while seemingly neutral, carry symbolic meanings linked to gender stereotypes. Subsequently, after cataloguing the images generated by AI based on these searches, they will be subjected to the Documentary Method of Interpretation, located within empirical social research and refined by the German sociologist Ralf Bohnsack.The procedural method is monographic, employing bibliographical and documentary research techniques. Ultimately, the study will propose guidelines for the improvement of Public Policies aimed at fostering a non-discriminatory Artificial Intelligence, which may contribute to the Brazilian Artificial Intelligence Plan. The proposed theme is appropriate to the concentration area of Social Rights and Public Policies and to the research line Diversity and Public Policies of the Graduate Program in Law at the University of Santa Cruz do Sul, and it is aligned with the research developed by the advisor, Professor Doctor Marli Marlene Moraes da Costa.