ALGORITHMIC BIAS AND SOCIAL INEQUALITY IN AI DECISION-MAKING SYSTEMS FROM A SOCIOLOGICAL PERSPECTIVE

Abstract

The rapid adoption of Artificial Intelligence (AI) in decision-making systems has brought new efficiency gains but also intensified debates about fairness, equity, and social justice. From a sociological standpoint, algorithmic bias is not merely a technical anomaly but a structural reflection of pre-existing social inequalities embedded in historical data, institutional practices, and cultural norms. This study investigates how algorithmic systems in domains such as criminal justice, health care, housing, and employment perpetuate and sometimes exacerbate inequality. Through an extensive review of empirical studies, this work examines mechanisms of bias across the AI lifecycle and the sociological theories that explain them, such as intersectionality, critical race theory, and social stratification, and presents case-based statistical evidence, including disparities in facial recognition accuracy, credit scoring, and risk assessment tools. The study also outlines methodological approaches for studying AI bias sociologically, discusses results through both quantitative metrics and qualitative interpretations, and suggests future perspectives for designing systems that are substantively fair. The findings emphasize that addressing algorithmic bias requires an interdisciplinary approach that bridges sociology, computer science, and policy.