ALGORITHMS AND DISCRIMINATION: TENSIONS WITH ANTI-DISCRIMINATION LAW IN THE CONTEXT OF DIGITALIZATION

Abstract

This article analyzes the legal challenges arising from the increasing use of artificial intelligence in decision-making processes, with a focus on the reproduction and amplification of structural discrimination by algorithmic systems. It departs from the premise that automated decision-making is not neutral, as it may incorporate and perpetuate historical asymmetries embedded in training data and modeling choices. Through bibliographic and documentary research, the article examines the legal concept of direct and indirect discrimination, the mechanisms through which algorithmic biases operate, and concrete experiences of public administration digitalization that have produced discriminatory outcomes. It further analyzes national and international regulatory frameworks aimed at artificial intelligence governance and the mitigation of discriminatory risks, with particular attention to risk-based and fundamental rights-based approaches. It concludes that the development of algorithmic systems compatible with the constitutional order and international human rights commitments requires the implementation of robust mechanisms of transparency, auditability, accountability, and oversight, capable of ensuring substantive equality and non-discrimination in the digital environment.