Racial bias in artificial intelligence systems challenges for civil liability and the need for risk parameterization
Main Article Content
Abstract
This article addresses the challenges of civil liability in cases involving wrongdoings related to biased artificial intelligence systems concerning racial matters. It explores the possibility of an algorithm being deemed racist—which is not confirmed in the conclusions—and how this can lead to harm due to the omission of proper data curation, resulting in algorithmic bias. Furthermore, the need for normative parameterization based on risk gradation, in line with the ongoing discussions in the European Union, is discussed. The research adopts a deductive method, relying on bibliographic support and revisiting theoretical concepts. The objective is to reach a conclusion on this problematic issue, considering the preservation of human rights and the challenges presented by the evolution of algorithms and artificial intelligence systems.