Crafting Good Indicators

Human-Machine Entanglements in Brazil's 2022 Population Census

Bibliographic Data

ID4380229
AuthorsM Kopper (0000-0003-0327-7681, corresponding author), Ulisses Corrêa Duarte
Year2024
Volume42
Issue1
Pages42-60
Publication date2024-03-01
Peer ReviewedYes
Open AccessYes
TypeARTICLE
VenueThe Cambridge Journal of Anthropology (JOURNAL)
Journal identifiersISSN: 0305-7674 • E-ISSN: 2047-7716
PublisherBerghahn Books (PUBLISHER • GB)
DOI10.3167/cja.2024.420104
OpenAlexW4396836885
LanguageEN
Citations received1
References cited18

Big data analytics have radically transformed data collection protocols in population censuses. Yet despite the unprecedented degree of automation these technologies afford, much of the critical work that goes into making data 'actionable' still hinges on the ethical labour of data operators in these novel human-machine settings. By following training activities and conducting interviews with technologists who worked on Brazil's 2022 population census, this article traces the workings of a fraud-detecting system designed to reduce costs and improve data quality and collection. Our ethnography identifies two competing modes of truth-making, which we term 'probabilistic' and 'performative', whereby numbers and census-takers 'tame' each other. Tracking human-machine data entanglements on the fringes of calculation centres helps unpack how futurities are affectively negotiated and woven into the political fabric of these political technologies

Census and Population Estimation · Ethics and Social Impacts of AI · Race, Genetics, and Society · Geology

Tools

Open DOIOpen Access