FACIAL RECOGNITION THROUGH ARTIFICIAL INTELLIGENCE AND THE PRESUMPTION OF INNOCENCE PRINCIPLE

investigations into brazilian and comparative criminal proceedings

Abstract

In this article, we attempt to establish shared aspects between facial recognition carried out by artificial intelligence for the purposes of predictive policing and recognition of people used as evidence in the Criminal Procedure Code. Although facial recognition through artificial intelligence is presumably scientifically impartial, it ends up reproducing the weaknesses of recognition carried out by people because it requires a pattern to be reproduced, inductively. Thus, this article aims at discussing the problems of facial recognition through artificial intelligence of a normative and epistemological nature. Other implications of algorithmic facial recognition are also worked on, such as privacy and the presumption of innocence principle, constitutionally designed and not yet implemented. Lastly, the present study presents how algorithmic facial recognition has been dealt with in Brazil and other countries around the world.