AUTOMATIZOVANA DISKRIMINACIJA

Abstract

The principle of equality represents a continuous line of development of the modern state community, while at the same time posing a challenge for an adequate response by the legal order. Confronting the achievements of new generations of artificial intelligence entails an obligation to formulate carefully considered and timely legal solutions for the protection of human rights. The risks arising from the inadequate application of artificial intelligence tools have been widely noted in legal scholarship, with particular emphasis on the potential of artificial intelligence to generate discrimination. The automation of processing large volumes of diverse data produces outcomes that carry a risk of violating the prohibition of discrimination, which is not easily proven. The non-transparency of the technological foundations of models embedded in systems renders artificial intelligence an insufficiently reliable instrument, one that threatens to limit and endanger the exercise of human rights. Cases from legal practice reveal the dangers faced by individuals in contemporary society and highlight the shortcomings of existing legal solutions and their application in the tense relationship between the prohibition of discrimination and the business interests of actors in the artificial intelligence market. There is hardly any individual human right that is not potentially at risk, yet the principle of equality appears to be increasingly vulnerable in the era of artificial intelligence. This paper examines general issues of artificial intelligence as a potential source of discrimination, insisting on the need for synchronized activity among multiple disciplines, with legal science occupying a particularly important place. Its responsibility toward the individual and society is becoming ever more pronounced, given the expectations placed on the legal order and on the legal positioning of artificial intelligence within the human rights system.