DISCRIMINATION OF ARTIFICIAL INTELLIGENCE ALGORITHMS:

AN ANALYSIS FROM THE VIEWPOINT OF FUNDAMENTAL RIGHTS

Resumo

Introduction: Innovations in fundamental rights are regularly preceded by their violation. Thus, innovations in fundamental rights and guarantees stem from social grievances, which the law attributes to the quality of violation. This legally immanent perspective is based on social and technological development. Using the 1988 Federal Constitution of the Republic of Brazil as a legal sieve and the United Nations Universal Declaration of Human Rights, everyone is a holder of proclaimed rights and freedoms, without distinction of any kind, whether of race, skin color, sex, language, religion, political or other conviction, national or social origin.  Discrimination in this context arises as a disadvantage and degradation of individuals or groups, through prejudices and conscious emotional associations, sometimes also ruled by the unconscious, a social phenomenon that contradicts the basic democratic principles and, therefore, a socially and morally forbidden policy. Methodology: This is a legal research, carried out by means of the qualitative-inductive method, using bibliographic research (various texts) and document analysis (laws and similar) as a technique. Development/Discussion: Approaching the discussion from a technological point of view, discrimination is not expected from machines and algorithms, since they are seen as neutral and only perform arithmetic operations, following the source code instruction. The touch point, is that algorithmic discrimination is becoming a major challenge and problem to fundamental rights, with the exponential increase of users in cyberspace. The technological-digital world is more and more oriented to the singular individual, addressing the personal characteristics of the latter, dissolving the very uniqueness of users. The automation in this line, occupies an important space, because the information technology systems are subjected to and make complex decisions autonomously, thus reducing the human supervisory activity. Conclusion: The driving factor in the adoption of algorithmic systems is their ability to process huge amounts of data (big data) and correlate numbers of data sets, coupling the unique methods of machine learning to statistical models (artificial intelligence - AI), however, this complexity and abstraction pervade an autonomous model of inference of opaque processes and do not provide clear explanations about the decisions that are made by artificial intelligence algorithms. This lack of transparency risks undermining the supervision and responsibility of the programmers, representing a risk in decision making, directly implying the violation of human rights.