ARTIFICIAL INTELLIGENCE SYSTEMS AND IMPACT ASSESSMENTS FOR HUMAN RIGHTS
Abstract
This paper, based on an exploratory approach, conducts an analysis of the functioning of artificial intelligence systems, questioning the impacts of the development and use of such systems on human rights. In this sense, it explores the discriminatory impacts for groups historically marked by vulnerabilities, seeking to demonstrate that the use of artificial intelligence systems can deepen already existing inequalities and violations, impacting such groups differently. At the normative level, we analyze the strategies for regulating artificial intelligence, focusing on the development of impact assessments from an approach centered on human rights. In conclusion, it is pointed out that, although impact assessments are important instruments for accountability and the formation of a governance structure for artificial intelligence systems, they must observe the existence of unacceptable risks, otherwise they may legitimize practices that violate human rights.