/ HCI & AI RESEARCHER

Building
Human-Centered
AI Systems

Kyusik Kim Kyusik Kim

I research Human-AI Interaction, designing AI agents that genuinely serve people. My work spans value alignment, fairness, societal impacts, and social simulation of AI systems.

Postdoctoral Researcher
at Seoul National University AI Institute.

4CHI
6ACL
4SIGIR
1EMNLP
18+Total

Understanding the Societal Impacts of AI

My research lies at the intersection of Human-AI Interaction, Value Alignment, and Fairness. I investigate how AI systems affect human decision-making, social dynamics, and cognitive processes—and how we can design them to be more equitable and aligned with human values.

I am particularly interested in understanding biases in multimodal AI systems, designing AI agents that support meaningful human experiences, and building simulation frameworks to study AI's societal effects at scale.

Human-AI Interaction

AI agents, co-viewing, shopping, and collaborative experiences

Value Alignment & Fairness

Bias in AI hiring, sycophancy, cross-cultural evaluation

Simulation & Benchmarking

LLM decision-making, clinical benchmarks, social simulations

Selected Publications

View All 17+ Publications →

Research Trajectory

Human-AI Interaction

AI agents for sports, cinema, shopping, and conversation

7 papers

Fairness & Bias

Cross-modal bias, halo effects, cultural asymmetry

5 papers

Value Alignment

Sycophancy, selective exposure, decision-making

3 papers

Social Simulation

LLM-based agent simulation, social polarization, policy testing

Current & Future
Explore Research Map →

Recognition

2026

Outstanding Dissertation Award

Graduate School of Convergence Science and Technology, Seoul National University

2025

Yulchon AI Young Researcher Scholarship

Selected as one of 10 recipients for outstanding "AI + X" interdisciplinary research

2024

Google Travel & Conference Grants

Student Travel Grants Program for research conference attendance

2024

BK-21+ Excellence Award

Selected as excellent graduate student in Brain Korea 21 program

2020–25

BK-21+ Scholarship

Government scholarship based on publication potential (5 years)

View Full CV & Awards →

Social Conflict Simulation with LLM-Based Agents

2026 – 2027 SNU AI Institute NRF Global Agenda

Toward Mitigating Social Polarization in Korea

Contributing to a project that builds social conflict simulation and policy feedback systems using LLM-based agents. My work focuses on constructing a Korean agent library (200+ agents reflecting demographics, political orientation, and cultural context) and validating polarization representation through 1:1 and small group debate simulations with quantitative metrics and human evaluation.

01Korean Agent Library Construction (200+ agents across 7 data categories)
02Polarization Representation Validation via Debate Simulations

Let's Build Responsible AI Together

I'm always open to discussing research collaborations, especially in human-AI interaction, value alignment, and the societal impacts of AI systems.