Researching local government at scale with AI
Sabina Tomkins (University of Michigan)
Abstract
Local governments in the United States oversee budgets which amount to trillions of dollars. They are highly consequential sites of political participation and offer a rare opportunity for direct democracy. Yet, due to the vast number of them (roughly 10,000 different units of local government in the US), they have been difficult to study. This is changing as many local governments have begun to post recordings of their meetings online (for example on YouTube). Now, computational methods can be harnessed to study these meetings at scale. This talk will present a large-scale research project for collecting, processing and analyzing these records and discuss the kinds of research opportunities they afford. It will conclude with insights from existing research projects, for example around questions of what kinds of concerns the public bring forward in these meetings and whether these concerns reflect local or national interests.