Rethinking Racial Inequality in Artificial Intelligence: A Critical Reflection through Cybernetics

Abstract

In my second year at the University of Sussex, I wrote an essay for the AI, Culture and Society module exploring how algorithms and data collection practices can reproduce racial inequality, and what could be done to make AI systems fairer and more accountable. Drawing on Frantz Fanon, Simone Browne and Shoshana Zuboff, I connected discriminatory outcomes to historical data, institutional decisions and commercial priorities. Returning to that work, I question whether its broad account of bias sufficiently explained the relationships between these different sources of harm. This reflection develops that question through cybernetics, which examines how the consequences of an action return as information that shapes later action. Predictive policing provides a focused example, as decisions on where officers patrol influence the records used to guide subsequent deployment. Although my original essay recognised this feedback, I had not consistently followed its implications either for the structure of my argument or the remedies I proposed. I reconsider the essay’s strengths, its assumption of a recoverable neutrality and the reassurance offered by terms such as transparency and accountability. The reflection leads towards a more focused research practice, attentive to how institutional decisions produce the evidence used to justify them, and to whether those affected possess the power to challenge that process.