Afterlives of Authority: An Ethnography of Fire Prediction, Social Order, and Technocracy in Indonesia

Abstract

This dissertation is an ethnographic study about how “data-driven” technologies are transforming environmental governance in Indonesia. It documents the increasing application of data science techniques by government research agencies to predict peatland fires, and how these techniques are reorganizing the work and livelihoods of technocrats. I argue that the promise of data-driven environmental governance is to create order out of uncertainty in rural Indonesia. This promise to quantify and control uncertainty has expanded the reach of the police and military state in Indonesia, undermining democracy in Indonesia that had largely been considered successful by public commentators internationally.

My findings are based on three years of ethnographic fieldwork and interviews with two government engineering and science research institutions, the National Institute of Aeronautics and Space and the Agency for Application and Assessment of Technology in Indonesia, and with tech corporations and federal science agencies from the United States. I use two case studies to support my argument that peatland fire prediction efforts have motivated the expansion of the military state in Indonesia: 1) the redesign of fire detection algorithms and methodologies developed in collaboration between the National Institute of Aeronautics and Space and North American federal science agencies; and 2) the design and deployment of a peatland fire prediction system using agile software development strategies and data science by the Agency for the Application and Assessment of Technology and the contracted tech corporation IBM.

The analyses that follow in this dissertation are inspired by postcolonial science and technology studies (STS) and critical data studies. As part of my dissertation, I show how what was framed as an opportunity to both resolve the failures of technocratic knowledge production and efficiently monitor the environment for endemic fire risk produced a system of automated surveillance upon poor smallholder farmers. Not only were scientists becoming para-law enforcers through the use of predictive technologies, but their datafied predictions rested on cultural stereotypes of farmers as uncivilized subjects who are unable to manage peatland. A aim of the dissertation is to answer the question: how are political responsibility and regulation in Indonesia outsourced to code and predictive systems? What kind of state ideologies and practices are reinforced or transformed as a result? In asking these questions, I situate my research amid literature in critical data studies, examining how it is that algorithms “enact governance” and what this means for a newly democratic country.