Computer Science Seminar: Emma Lurie (University of Pennsylvania)

Speaker:  Emma Lurie (University of Pennsylvania)

Title: Pursuing Accountable AI for High-Quality Democratic Discourse

The seminar will be available for in-person and Zoom participation. To participate online, please email inquiry-cs@barnard.edu to receive the Zoom link.

AI systems increasingly mediate access to information and shape public discourse, yet how they cause harm—and who is responsible—remains contested. In this talk, I present two projects that address these questions by examining AI governance in contexts important to democratic participation. First, I analyze how search engines can surface misleading information about ballot propositions during elections. Through a sociotechnical audit, I reveal how platform design choices, commercial incentives, and election officials’ interventions jointly shape voter access to information about democracy. Second, I discuss ongoing work monitoring how large language models moderate content about social issues over time, uncovering patterns in how these systems handle political speech and revealing inconsistencies that raise concerns for a democratic society. Together, these projects demonstrate an interdisciplinary approach to responsible AI research: examining not just technical failures, but how harms emerge from interactions between algorithms, institutional actors, and regulatory frameworks. I discuss how this work bridges computer science, law, and social science to identify intervention points for more accountable AI systems.

Emma Lurie is a postdoctoral researcher at the University of Pennsylvania conducting civil liberty-focused audits of AI systems. As a public interest technology researcher and lawyer, she develops and deploys empirical methods to assess AI systems in high-stakes contexts like election information and government surveillance. Her research bridges computer science, law, and information science to examine how AI systems produce harm and how responsibility is allocated across legal and technical domains. Her work has been published at venues including FAccT, AIES, and New Media & Society. She holds a PhD in Information Science from UC Berkeley and a JD from Stanford Law School.