The single most influential historical study of how British colonial administration produced "knowledge" — censuses, surveys, ethnographies, maps — that simultaneously described and constituted the categories it claimed to measure. Foundational reading for anyone working with Indian official data.
Deep Dive · Data & Society
Data, Power, and the Global South
Who counts, who is counted, and who decides — a critical syllabus on data politics in the developing world.
Data Justice
AI
11 readings
IM
ImpactMojo Editorial
Curated by the ImpactMojo team
A house-curated reading list assembled from the materials we use when teaching data ethics and AI to development practitioners. We're looking for an invited curator from critical data studies or the Global South tech-policy world to take this further.
House Pick
Editor's Note
Most data-and-development teaching focuses on what you can do with data: dashboards, machine learning, RCT analysis. Most data-and-power thinking focuses on what data does to people: surveillance, exclusion, automation of bias. The first body of work is overwhelmingly Northern in origin and ambient assumptions; the second is increasingly being written from the Global South. Practitioners need both.
This list is built around that gap. It begins with histories — because the systems we now call "data infrastructure" are continuous with colonial census-making and enumerative statecraft — then moves through the contemporary critical literature on data justice and AI bias, and ends with the most useful Global South voices working on what data and AI mean for everyone who is not the average user assumed by the technology stack.
Short on time? Read D'Ignazio & Klein and Arora. Want one essay that captures the political stakes? Read Linnet Taylor on data justice.
Section 01
Histories of Counting
The colonial and post-colonial origins of the data infrastructures the development sector now treats as neutral.
The most cited single articulation of "data justice" as a framework. Taylor sets out three pillars — (in)visibility, (dis)engagement, and non-discrimination — that have become the standard analytical lens for thinking about data and power outside the consumer-rights frame.
A rigorous philosophical case for the partial, situated character of all scientific knowledge. Read for the methodological grounding it provides for taking "Southern" data critique seriously rather than treating it as an exotic supplement to mainstream methods.
Section 02
Data Justice
The contemporary literature on what equitable data practice would look like — and on the harms when it doesn't exist.
The clearest book-length treatment of data ethics for practitioners. Each chapter takes a principle (examine power, challenge power, elevate emotion, etc.) and works through it with case studies. Especially useful for evaluation and dataviz teams that want to move beyond a "neutral data" frame without losing methodological rigour.
A working classic on what happens when welfare states automate. Eubanks's case studies — the Indiana welfare system, the LA homelessness algorithm, the Allegheny child-welfare model — show that automation rarely simplifies the safety net for the people most reliant on it. Read alongside any work on India's welfare digitisation.
Benjamin's notion of the "New Jim Code" — that contemporary technology can reproduce and disguise discriminatory outcomes more effectively than overtly racist policy — has become one of the field-defining frames. Although the empirical context is the US, the argument transposes directly to caste, religion, and gender in the South Asian context.
Section 03
AI and the Global South
What language models, computer vision, and prediction systems mean for the four billion people on whom they were not trained.
A field-defining account of what internet and AI use look like outside the Northern consumer-tech imagination. Arora's research is grounded in extensive fieldwork across the Global South and reframes "the next billion users" as agents with their own purposes, not as a development problem.
The empirical paper that brought "algorithmic bias" out of the academy and into mainstream public debate. The methodology — auditing commercial face recognition systems for differential accuracy by skin tone and gender — is the template for a generation of subsequent AI fairness research.
The single most useful contemporary news source on technology in the Global South. Long-form reporting on platform labour, surveillance, fintech, and AI from places that mainstream tech press ignores. Treat the archive as a perpetually updated working syllabus.
Section 04
Decolonising Data
South Asian voices on the politics of data, surveillance, and digital identity at home.
An edited volume that collects the most rigorous critiques of India's biometric identity system — exclusions in welfare, surveillance risks, and the gap between project promise and audited delivery. The empirical chapters are the strongest single-source reference for evaluators reasoning about welfare digitisation.
The most consistent source of careful, technically literate Indian research on internet policy, surveillance, and data protection. The CIS reports on the Personal Data Protection Bill, on Aadhaar, and on platform liability are required reading for anyone working at the intersection of tech and policy in India.
Suggested citation
ImpactMojo Editorial (2026). "Data, Power, and the Global South." ImpactMojo Deep Dives. Retrieved from https://impactmojo.in/DeepDives/data-power-global-south.html
Want to curate a Deep Dive?
If you teach, research, or practice in development and have a reading list worth sharing — pitch us.
Pitch a Deep Dive →