Discrimination in the Digital Workplace: New Manifestations and Responses

Abstract

Advances in AI tools and algorithmic systems have significantly reshaped the modern workplace, presenting new challenges to achieving equality and non-discrimination in workplace relations. Machine-learning algorithms are being adopted in recruitment, employee assessment, work assignment, performance monitoring, and termination. Although these technologies enable greater efficiency and consistency, they also create risks of algorithmic discrimination, which may occur unintentionally and be difficult to detect and challenge legally. The aim of this study is to examine the extent to which existing international legal instruments effectively regulate employment relations in the context of algorithmic discrimination and to identify gaps in the law that should be addressed through further development. The research adopts a doctrinal and socio-legal approach, drawing on doctrinal legal analysis of international legal instruments, qualitative thematic analysis of academic literature, comparative legal analysis, and a study of judicial and quasi -judicial practice concerning employment decisions made with the assistance of AI and platform work. The results show that international standards provide a valuable normative baseline for safeguarding workers against discrimination but are currently fragmented and inadequately tailored to algorithmic decision-making. The study introduces an original six-component typology of algorithmic discrimination, covering historical-structural, data-selection and measurement, proxy-based, design-and-policy, deployment-and-governance, and emergent-group discrimination. It also establishes a three-part test for algorithmic discrimination, comprising the use of an AI-assisted decision, a differential or disproportionate impact on employees, and a lack of transparency, explainability and meaningful human review.