Algorithmic decision-making and the right to non-discrimination: regulation and bias mitigation in the Era of artificial intelligence

Abstract

A fundamental objective of the Federative Republic of Brazil is to promote the well-being of all, without prejudice of origin, race, sex, color, age, or any other forms of discrimination, as ensured by Section IV, Article 3 of the 1988 Federal Constitution. In parallel, the rapid development of digital innovation and the increase of disruptive technologies in people's lives, such as artificial intelligence, is well known. However, social problems and violations of fundamental rights are part of everyday life in Brazil, and new technologies can amplify these threats, encouraging social discrimination through the incorporation of discriminatory biases. In light of this issue, this paper engages in a discussion about the necessary regulation of artificial intelligence in Brazil, conducting an analysis of Bill 2.338/2023, approved by the Federal Senate on December 10, 2024, in order to investigate if and how this AI regulation project protects the fundamental right to non-discrimination, which is threatened by the unregulated use of artificial intelligence systems. Additionally, this study also conducts a comparative analysis with the pioneering AI regulation legislation approved by the European Union, examining how the normative provisions aimed at protecting the fundamental right to non-discrimination were established there.