A promoção da segurança pública brasileira através das tecnologias de reconhecimento facial: a necessidade de sua regulamentação no horizonte dos direitos fundamentais

Abstract

Artificial intelligence technologies are already a part of many of the everyday chores of human life. In the last decade, there’s been a noticeable public investment in tests and in the use of automated facial recognition systems to promote brazilian public security. In this context, the creation of a specific regulation of the topic is needed and raises questions about what its outlines should be. Furthermore, the aim of this research is to comprehend how this technology works and interpret the underlying aspects of its use, beyond its technical aspects; to study the use of this tool in the promotion of the brazilian public security and its impacts on the fundamental rights; and to map out brazilian laws and bills on the topic, comparing them to foreign regulamentations. To this end, the hermeneutic method, the monographic procedural method and a documental research are used. The conclusion is that artificial intelligence models present flaws and biases that lead to the reproduction of racial prejudice and social stigmas. In addition, the automated facial recognition can be used as a mass surveillance tool, which constitutes a violation of fundamental rights and civil liberties. Its use in the brazilian context results in the perpetuation and possible intensification of criminal selectivity and mass encarceration. Lastly, the lack of transparency on the part of the governamental entities that use this tool makes it impossible to measure its efficiency, but the avaliable data points to the negative impacts outweighting the benefits of its use. Thus, if the future legislation, that is yet to be created, doesn’t ban this technology, it should attend to the constitutional rights and protect their effectiveness, as well as instrumentalize the general principles of data protection that are listed by the Lei Geral de Proteção de Dados and apply the precautionary principle in the redaction of the normative text.