Do Won Kim

I am a Ph.D. candidate in Information at the University of Maryland, advised by Giovanni Luca Ciampaglia and Cody Buntain. I am also a student affiliate at TRAILS (The Institute for Trustworthy AI in Law and Society).

I view platform architectures—particularly recommendation systems—as institutions that govern how attention is distributed, which voices are amplified, and how people interact online. Using computational and experimental methods, I intervene in these systems to understand how they influence users’ information exposure, behaviors, and attitudes.

My dissertation addresses the gap between what users say they prefer (stated preferences) and what their behaviors suggest they prefer (revealed preferences). Social media algorithms infer user preferences from engagement behaviors, even when those behaviors don’t reflect what users actually value, often amplifying content that is attention-grabbing but misaligned with users’ true values. How can we design social media feed algorithms that better align with what users consciously value? And what effects might such value-aligned systems have on individuals and society? To address these questions, I develop a stated preference-based recommendation algorithm that leverages LLMs to re-rank content based on users’ explicitly expressed values, and test its effects on both individual- and system-level outcomes through RCTs and large-scale simulations.

Beyond the dissertation, I have conducted a digital field experiment on Twitter/X to examine how muting low-quality sources influences users’ information consumption and downstream political attitudes. I was also part of a finalist team in the Prosocial Ranking Challenge, where we designed a recommendation algorithm aimed at fostering constructive, cross-partisan dialogue online. In another collaborative project, I study how real-time political conversations can reduce affective polarization using LLM-embedded experiments.

Building on these lines of work, I am expanding my research into the domain of human-AI interaction with a focus on civic applications. I explore how AI can be leveraged to support more prosocial and pro-democratic engagement online, such as by supporting constructive political dialogue or helping bridge partisan divides. Methodologically, I am also interested in integrating AI into social science experiments, including AI-augmented survey designs that simulate dynamic conversations or deliberative settings at scale.

News

[April. 2025]
1 talk + 2 posters accepted for IC2S2 2025!
[Jan. 2025]
Paper accepted for the IPSA World Congress of Political Science!
[Dec. 2024]
Passed the milestone and advanced to PhD candidacy!
[Nov. 2024]
Awarded the College Information Studies Alumni Chapter Scholarship ($1,000).
[Sept. 2024]
Awarded the Humane Studies Fellowship from the Institute for Humane Studies ($4,000)
[Aug. 2024]
Awarded the Doctoral Students Research Awards ($1,500)
[July. 2024]
Awarded the Jacob K. Goldhaber Travel Grant ($250) for IC2S2!
[June. 2024]
Had fun participating in A Short First Course in Network Science
[May. 2024]
🥳 Our team was selected as one of nine finalists of the Prosocial Ranking Challenge, a worldwide crowdsourcing challenge organized by the Center for Human-Compatible AI at UC Berkeley!
[April. 2024]
Accepted a poster presentation for IC2S2 2024!
[April. 2024]
Served as a Program Committee member for the International Workshop on Cyber Social Threats (CySoc 2024) at the 2024 International Conference on Web and Social Media (ICWSM).
[April. 2024]
Became an affiliated PhD student at the Institute for Trustworthy AI in Law and Society (TRAILS).
[Mar. 2024]
Reviewed a paper (Behaviour & Information Technology).
[Feb. 2024]
Reviewed ICWSM 2024 papers.
[May. 2023]
🥳 I am thrilled to have received the Dean's Fellowship during my first year of doctoral studies at UMD!