ALGORITHMIC BIAS AND THE JUDICIARY: AN INVESTIGATION OF EXISTING LEGAL MECHANISMS

Abstract

The present research aims to investigate the existence of effective legal mechanisms to curb the influence of ideology in programs that use AI in the judicial field. As specific objectives, it brings the examination of the theoretical foundations of algorithmic bias, the analysis of the use of AI in the Judiciary and, finally, the investigation of possible mechanisms of filtering algorithmic bias. Regarding methodology, it is a qualitative research, considering that the proposed discussion needs contextualization and cannot be reduced to statistical data. The hypothetical-deductive method was used through the formulation of the hypothesis: "The legal mechanisms in force are sufficient to curb the influence of ideologies and values in the programming and operation of AI". It is concluded that, in the specific case, the existing legal mechanisms are insufficient. Artificial intelligence has the potential to increase the speed and effectiveness of procedural analysis, but it needs to undergo rigorous tests to rule out possible discriminatory traits. AI should be complementary to judicial decision-making, not an essential element of it.