INTRODUCTION TO CODED DISCRIMINATION

Abstract

Discrimination is defined as an unjust distinction based on personal characteristics and is prohibited by all relevant international documents related to the protection of human rights. The definition, forms, prevalence, and protection mechanisms against discrimination have garnered the attention of legal theorists, policymakers, and practitioners alike. While it may seem that there is nihil novum sub sole regarding this topic, the rapid expansion of artificial intelligence in various sectors of society brings forth a myriad of new questions. Artificial intelligence systems, particularly those based on machine learning, are prone to creating what is known as technical bias, which can further evolve into "coded discrimination." This term refers to discrimination that arises from the decisions made by computer programs operating based on predefined parameters. This paper aims to analyze the concept of "coded discrimination," present the relevant international framework for regulating artificial intelligence—focusing on provisions related to human rights and discrimination—and examine the applicability of domestic anti-discrimination regulations to instances of discrimination stemming from the use of artificial intelligence in the Republic of Serbia.