Organizational decision systems increasingly mediate recruitment, evaluation, allocation, scheduling, monitoring, and other consequential workplace processes. Debate about workplace inequality, however, often treats algorithmic bias as a technical property of models rather than as an organizational process involving data, objectives, authority, participation, work design, and accountability. This Critical Review Article evaluates that broader systems problem. It synthesizes recent peer-reviewed research on algorithmic control, automated human resource management, worker voice, surveillance, burden distribution, professional judgment, and governance. The review distinguishes statistical or technical discrimination from perceived fairness, voice opportunity from consequential influence, automation from augmentation, technological capability from legitimate authority, and algorithmic opacity from enacted accountability. Across the evidence, unequal outcomes can arise through several pathways: optimization may interact with unequal environments; workers may be exposed to decision systems they cannot meaningfully shape; errors and adaptation costs may be distributed unevenly; and responsibility may become fragmented across human and technical actors. Yet the evidence is heterogeneous, and platform-work findings, laboratory judgments, conceptual frameworks, and organizational cases cannot be treated as interchangeable. The article therefore proposes a Systems Model of Inequality Production as an integrative synthesis rather than a validated causal model. The model identifies linked decision layers, four inequality-producing mechanisms, feedback processes, and contextual gates that require direct longitudinal and multilevel testing. The principal implication is that organizational inequality cannot be evaluated adequately by model accuracy or fairness metrics alone.
