Regulation of algorithmic discrimination in international law: structural assessment and effectiveness of individual legal remedies from a human rights perspective

Abstract

Introduction. Due to rapid technological development, artificial intelligence systems (AI) are increasingly used in various social and legal domains, such as recruitment, judicial decisions, and credit scoring. Although these technologies may appear neutral, the data they rely on and their learning processes often replicate and reinforce pre-existing structural biases, resulting in both direct and indirect discrimination against certain social groups.Purpose. To analyze the issue of algorithmic discrimination arising in the decision-making processes of AI systems from the perspective of international law and human rights.Theoretical Basis. The article explores the legal and technical dimensions of algorithmic discrimination. Legally, such discrimination does not align with traditional concepts, as the focus shifts from “intent” to “impact.” This complicates legal assessment and the burden of proof. Technically, many algorithms function as “black boxes,” meaning their decision-making processes are opaque, making legal intervention more difficult.Results and Conclusions. It is argued that the current international human rights framework – including instruments such as the ICCPR and ICERD – is insufficient to address this issue. These normative gaps are due not only to technical complexity but also to weak state compliance and the absence of specific legal standards.Individual legal protection mechanisms – such as the European Court of Human Rights (ECtHR), UN Treaty Bodies, and the Inter-American Court of Human Rights – are currently ill-equipped to respond effectively due to limited technological expertise and procedural barriers.The article highlights the European Union’s relatively advanced approach through instruments like the AI Act and GDPR, while pointing out the significant gaps that remain in the United States and Global South countries. The article concludes with recommendations, including strengthening the international legal framework, aligning national legislation, establishing specialized institutions, and mandating algorithmic impact assessment.