Predictive Algorithms and Social Inequality: A Sociological Analysis of Bias, Governance, and Digital Surveillance

Abstract

The rapid expansion of predictive algorithms across social, economic, and governmental systems has intensified concerns about the reproduction of inequality and the emergence of new forms of algorithmic governance. This paper examines how predictive technologies—ranging from risk-assessment tools and automated hiring systems to credit-scoring platforms and predictive policing—systematically shape opportunities, access, and social outcomes. Drawing on insights from sociology, critical data studies, and surveillance theory, the study analyses the mechanisms through which algorithmic systems encode historical biases, amplify structural disadvantages, and normalise surveillance as a mode of social control. The paper argues that predictive algorithms operate within unequal data infrastructures that disproportionately disadvantage marginalised groups, reinforcing patterns of racialised, gendered, and class-based exclusion. Moreover, the opacity of algorithmic decision-making and the rise of automated governance shift power away from public accountability and towards computational forms of authority controlled by states and corporations. By situating predictive technologies within broader socio-political contexts, this paper highlights the urgent need for transparent modelling practices, anti-bias regulation, and equitable data governance frameworks. The findings contribute to ongoing debates about digital injustice and provide a sociological foundation for understanding how algorithmic systems reshape power, surveillance, and inequality in the digital age.