Data Justice & Algorithmic Harm

Eubanks documents how automated systems in public services — from child welfare algorithms to predictive policing — systematically disadvantage poor and working-class communities. Her case studies from Pennsylvania, Indiana, and Los Angeles show that "objective" algorithmic scoring often encodes existing inequalities, making their consequences harder to contest.

Noble analyses the racial biases embedded in commercial search algorithms, arguing that these systems function as instruments of power that reproduce white supremacist ideas through the apparently neutral operation of relevance ranking. Essential for understanding how statistical models trained on historical data perpetuate structural discrimination.

O'Neil coined the term "weapon of math destruction" for models that are opaque, unaccountable, and operate at scale — creating feedback loops that entrench disadvantage. Her accessible survey covers teacher evaluation, credit scoring, recidivism prediction, and targeted advertising, providing an essential vocabulary for critical data analysis.

Drawing on feminist theory and data science practice, D'Ignazio and Klein argue that data work is never value-neutral. The book introduces seven principles — from "examine power" to "make labour visible" — that provide a practical framework for critical data literacy. Particularly relevant for the counting lives strand of this project.

Benjamin's "New Jim Code" concept names the way technology encodes racial hierarchy while projecting neutrality. By tracing the design choices behind facial recognition, predictive policing, and health algorithms, she shows that racism is baked into technical artefacts — a structural argument that complements Eubanks' case-study approach.

Zuboff develops the concept of "surveillance capitalism" — an economic logic that claims human experience as raw material for behavioural prediction products. While wide-ranging in scope, Part II on the instrumentation of everyday life is most directly relevant to poverty measurement contexts where administrative data collection creates asymmetric power relations.

Pasquale examines the opacity of financial and search algorithms from a legal perspective, arguing for accountability and transparency requirements. His policy-oriented analysis is a useful complement to the sociological and critical race theory approaches of other entries, and introduces concepts of algorithmic accountability that have shaped subsequent regulation.

Milan and Treré challenge the assumption that "big data" is a universal phenomenon, arguing that algorithmic systems are experienced differently in the Global South. Their framework of "data universalism" — the imposition of Northern data practices and concepts onto Southern contexts — is essential for situating poverty measurement within global power structures.