The Political Responses of Virtual Assistants

Abstract

What would a machine say about politics if it could speak? Because of virtual assistants, such as Amazon’s Alexa or Apple’s Siri, social scientists are now equipped to answer this question. Yet little is known about what these machines say about politics or why they say it, even though their political responses are potentially consequential given their widespread usage. In this article, I survey six virtual assistants and find that they typically have an above average knowledge of politics, an elementary understanding of important political concepts, and only a handful of opinions on important political issues. In follow-up interviews with tech workers, ranging from software engineers to executives, I use grounded theory to generate hypotheses about how companies craft responses to political questions. I find that tech companies are motivated in large part by economic considerations and that the future responses of the virtual assistants will likely create political echo chambers in which responses are highly customized to individual users. I conclude by discussing the implications of these findings for political polarization, democratic citizenship, and the growing prevalence of artificial intelligence in politics.