Understanding Institutions by Analyzing Interactions | Rob Voigt (October 14, 2026)

It is common to think about critical public institutions such as “the police” or “the justice system” as abstract monoliths. Ultimately, however, these institutions are in large part an aggregate of many person-to-person interactions, often primarily conducted in the medium of language. In my work, I use tools from computational linguistics and sociolinguistics to measure large-scale linguistic trends in institutional settings and quantify disparate impact along social divisions such as race. In this talk, I will discuss a line of work on police-community interaction using computational models to analyze police body-worn camera footage in this way, addressing both procedural justice considerations related to interpersonal treatment as well as concrete issues of policy compliance and constitutionality. I will also describe the early stages of related work applying similar methods to 911 calls. Speaker Rob Voigt Rob Voigt is a computational sociolinguist who applies contemporary natural-language processing methods to questions about the linguistic mechanisms of the social world. He is interested in natural language processing, social meaning, sociophonetics and interactional variation, Chinese and Japanese linguistics, gesture and embodiment, and applications of linguistic research to social good. He is particularly interested in using computational methods to understand the linguistic mechanisms of social problems in domains like policing, gun violence, immigration, and mental health. Voigt is an assistant professor at UC Davis Linguistics, where he directs the Linguistic Mechanisms Lab. More info: https://www.ischool.berkeley.edu/events/bellwether-lectures/2026/understanding-institutions-analyzing-interactions