Files

Abstract

This dissertation makes the case that emerging computational technologies—encompassing machine learning tools, the data that powers them in commercial settings, and the consumer services and devices they optimize—hold theoretical and methodological significance for the field of international relations. These technologies enrich our understanding of international politics in two ways. First, computational technologies hold consequences for understanding the contemporary international order, specifically because technology firms use them to shape digital information transmission. Second, scholars can use these same computational technologies to offer novel insights across diverse debates in the field. In the first chapter, I argue that large technology companies are critical actors in international politics due to their ownership of data collection and processing infrastructures. In short, digital trace data—the data generated by interactions with internet-connected devices—and the technologies that collect and process it allow technology firms to set the terms by which an immensely differentiated world is commonly understood. I draw on Susan Strange’s work on structural power to argue that technology firms shape the choice sets of other actors by mediating information environments. This permits these firms to acquire structural power, which they exercise in at least three ways. First, firms mediate the distribution of legibility derived from digital trace data, whereby they decide the terms under which information is distributed. Second, firms seek to maximize their profit by shaping individual interests, often affecting political outcomes. Finally, firms use structural power to perform state-like functions both domestically and internationally. The second chapter discusses the role of internet search engines in shaping digital information flows, a subject surprisingly neglected in international relations literature. Co-authored with Dr. Rochelle Terman, this chapter presents an audit of Google search engine result pages across international affairs topics, revealing three primary findings. First, we find substantial variation in the reach of ideological content, including state propaganda and material from transnational advocacy organizations. Second, search results strongly correlate with search language, suggesting that language is a primary factor mediating exposure to political information. Finally, we analyze search results related to the war in Ukraine generated both before and after the 2022 Russian invasion and find more pronounced geographic clustering in post-invasion results, especially among states in less common language groups. In the final chapter, I demonstrate the utility of computational technologies in providing novel insights into other debates in the field. I develop a word embedding methodology to quantify friendship and collective identity between states. By applying this methodology to a corpus of all speeches in the United States Congress from 1899-2017, I create biennial friendship and collective identity measures between the United States and 192 countries. I find that perceptions of friendship between the United States and other countries grew dramatically following World War II, while perceptions of collective identity have slowly grown over the past eight decades. Moreover, the correlates of collective identity evolved over the 20th century to match those that predict enmity, suggesting the existence of a shared identity among enemies in international politics.

Details

Actions

PDF

from
to
Export
Download Full History