Data Feminism and the Politics of Counting
Catherine D'Ignazio and Lauren F. Klein's Data Feminism is one of the most useful books for keeping AI governance grounded before the model appears. It asks what gets counted, who does the counting, who is missing, and whose power is made to look natural by a chart, database, or model.
The Book
Data Feminism was published by MIT Press in 2020 and is also available as an open-access edition. Its seven principles connect data science to intersectional feminism, power analysis, pluralism, context, labor, embodiment, and justice.
The book is practical rather than ornamental. It is about how datasets, dashboards, categories, and visualizations can either reproduce domination or make hidden conditions visible enough to challenge.
Data Work Is Power Work
The most important lesson is that data does not arrive from nowhere. It is collected by institutions, shaped by categories, cleaned by workers, limited by incentives, and interpreted through social assumptions. A missing field, a default category, or a convenient proxy can become an administrative fate.
Data Feminism is especially strong on the politics of absence. Missing data is not always a technical flaw. Sometimes absence reflects neglect, danger, privacy, exclusion, or refusal. Treating absence as mere noise can erase exactly the people a justice-oriented system should notice.
The AI-Age Reading
In the AI era, data feminism is not a niche correction to model culture. It is upstream safety work. Training data, benchmark data, evaluation data, red-team data, metadata, labels, and user telemetry all carry politics before they carry signal.
Generative AI intensifies the problem because data can be laundered through fluency. A model can summarize a biased archive in a neutral tone, generate confidence from an incomplete corpus, or turn a contested category into an apparently settled answer. The interface hides the politics of counting behind the smoothness of language.
The Site Reading
For this site, Data Feminism belongs at the foundation of any serious AI-governance shelf. It gives language for the work that happens before deployment: category design, dataset documentation, consent, labor visibility, community review, and the right to challenge a system's frame.
The practical instruction is simple: do not audit only the model. Audit the world the model was allowed to see. Then audit the reasons that world was recorded in that form.
Sources
- MIT Press, Data Feminism publisher page.
- Catherine D'Ignazio and Lauren F. Klein, Data Feminism open-access edition.
- Amazon, Data Feminism by Catherine D'Ignazio and Lauren F. Klein.
Book links are paid affiliate links. As an Amazon Associate I earn from qualifying purchases.