Human-Computer Interaction

Human-Computer Interaction (HCI) is an interdisciplinary field that explores the design, use, and impact of interactive computing technologies on individuals and society. Our research investigates how technology mediates human experiences in diverse social, emotional, and cultural contexts—from online teaching and mobile shopping to family communication and survivor advocacy. We examine the roles of interfaces, social presence, and user agency in shaping these experiences, and we design systems that respond to real-world needs such as facilitating long-distance care through telepresence and reducing computer vision syndrome through deployable interventions. By combining qualitative insights with system design, we aim to create technologies that are empathetic, inclusive, and grounded in lived experience.

Research Highlights
  • I feel being there, they feel being together: Exploring How Telepresence Robots Facilitate Long-Distance Family Communication (2024), J. Seo, H. Lim, B. Suh, and J. Lee, CHI 2024
    This qualitative in-situ study with eight families over two weeks reveals that telepresence robots—through physical embodiment and autonomy—enable immersive, natural, and dynamic long-distance family interactions across behaviors like multi-party communication, exploring the home, restoring routines, support, and joint activities. [Honorable Mention | link]

  • Investigating the Effects of a Real-time Student Monitoring Interface on Instructors’ Monitoring Practices in Online Teaching (2024), H. Lee, S. Park, E. H. Kim, J. Seo, H. Lim, and J. Lee, CHI 2024
    A controlled experiment with 20 instructors shows that real-time student engagement monitoring interfaces enhance monitoring quality and teaching effectiveness, but customizable interfaces are crucial to balance engagement cues’ benefits against cognitive load and psychological stress on instructors. [link]

  • The Power of Close Others: How Social Interactions Impact Older Adults’ Mobile Shopping Experience (2023), J. Seo, Y. E. Park, E. H. Kim, H. Lim, and J. Lee, CHI 2023
    The study finds that older adults’ mobile shopping experiences are shaped by three key social interactions—learning from, collaborating with, and assisting close others—which help build trust in the systems and support decision-making. [link]

  • Design Guidelines of a Computer-Based Intervention for Computer Vision Syndrome: Focus Group Study and Real-World Deployment (2021), Y. Hwang, D. Shin, J. Eun, B. Suh, and J. Lee, Journal of medical Internet research
    A 14-day deployment of the LiquidEye system identified key interface features—clear resting instructions, customizable break goals, positive feedback, mid-size pop-ups, and symptom reminders—that significantly improved engagement in managing computer vision syndrome, informing design guidelines for eye health tools. [link]

  • TalkingBoogie: Collaborative Mobile AAC System for Non-verbal Children with Developmental Disabilities and Their Caregivers (2020), D. Shin, J. Song, S. Song, J. Park, J. Lee, and S. Jun, CHI 2020
    This paper presents TalkingBoogie, comprising two mobile apps—TalkingBoogie-AAC (for caregiver–child communication) and TalkingBoogie-Coach (for caregiver collaboration)—which facilitate structured symbol layouts, guided observations, and balanced participation; a two-week study with 11 participants showed improved mutual understanding and reduced cognitive load among caregivers. [Honorable Mention | link]

Human-AI Interaction

Human-AI Interaction is a research area that focuses on designing and understanding interactions between humans and AI systems. It explores how different types of AI systems form relational identities with users, going beyond simple information delivery to investigate the potential of relational AI. The research emphasizes practical applications, including LLM-based techniques for understanding and optimizing ambiguous user utterances, AI-mediated information sharing in care environments, and intuitive communication design for older adults. Recent work also explores the use of large language models to simulate people’s mental models and create digital doppelgängers—AI agents that reflect users’ perspectives and relational patterns—to study the dynamics of social relationships in human-AI interactions.

Research Highlights
  • Letters from Future Self: Augmenting the Letter-Exchange Exercise with LLM-based Agents to Enhance Young Adults' Career Exploration (2025), Jeon, H., Yoon, S., Lee, K., Kim, S. H., Kim, E. H., Cho, S., Ko, Y., Yang, S., Dabbish, L., Zimmerman, J. and Kim, E.M., & Lim, H, CHI 2025
    This study enhanced a future-self letter-exchange exercise for career development by integrating LLM–based agents that simulate participants’ future selves. In a one-week experiment (N=36), AI-generated letters and chat-based interactions increased engagement compared to manual letter writing. [Best Paper | link]

  • SPeCtrum: A Grounded Framework for Multidimensional Identity Representation in LLM-Based Agent (2025), Lee, K., Kim, S.H., Lee, S., Eun, J., Ko, Y., Jeon, H., Kim, E.H., Cho, S., Yang, S., Kim, E.M. and Lim, H., NAACL 2025
    This study introduces SPeCtrum, a framework for constructing authentic LLM agent personas by integrating Social Identity, Personal Identity, and Personal Life Context, finding that while life context alone can enable basic identity simulation, combining all three dimensions produces more comprehensive and accurate representations for real-world individuals. [link]

  • Human-AI Interaction in Human Resource Management: Understanding Why Employees Resist Algorithmic Evaluation at Workplaces and How to Mitigate Burdens (2021), H. Park, D. Ahn, K. Hosanagar, and J. Lee, CHI 2021
    Through in-depth interviews with 21 employees, the study identifies six types of burdens—emotional, mental, bias, manipulation, privacy, and social—associated with algorithmic HR evaluation, and suggests design interventions like transparency, interpretability, and human oversight to mitigate them. [link]

  • Bot in the Bunch: Facilitating Group Chat Discussion by Improving Efficiency and Participation with a Chatbot (2020), S. Kim, J. Eun, C. Oh, B. Suh, and J. Lee, CHI 2020
    The paper introduces GroupfeedBot, a facilitator chatbot designed to manage discussion time, promote balanced participation, and organize opinions in group chats, which led to increased opinion diversity without affecting message quantity or quality. [link]

  • Comparing Data from Chatbot and Web Surveys: Effects of Platform and Conversational Style on Survey Response Quality (2019), S. Kim, J. Lee, and G. Gweon, CHI 2019
    This experimental study finds that using a chatbot—versus a traditional web survey—can affect response quality, with conversational style (formal vs. casual) playing a significant role in engagement and data quality. [link]

AI Journalism

AI Journalism explores how artificial intelligence is transforming journalism and media technologies in the digital era. Our work investigates AI-driven news generation, automated fact-checking, and personalized news recommendation systems to design new forms of media experiences. We have conducted long-term research on automated news writing (robot journalism) and developed large-scale datasets—over 100,000 training instances—for AI-powered fact-checking. Through these projects, we examine diverse applications of AI in journalism and explore how intelligent systems can enhance the accuracy, efficiency, and trustworthiness of news.

Research Highlights
  • Balancing Artificial Intelligence and Human Expertise: Ideal Fact-Checking Strategies for Hard and Soft News (2025), Y. Kim, and J. Lee, Journalism & Mass Communication Quarterly
    An online experiment in Korea and the U.S. shows AI-assisted fact-checking is more effective for objective hard news, while human-assisted checking excels for subjective soft news due to differing cognitive processing mechanisms. [link]

  • Understanding User Perception of Automated News Generation System (2020), C. Oh, J. Choi, S. Lee, S. Park, D. Kim, J. Song, D. Kim, J. Lee, and B. Suh, CHI 2020
    The study introduces NewsRobot, a prototype that automatically generated real-time news during the PyeongChang 2018 Winter Olympics, and finds that while users appreciated individualized and richly presented news formats, they perceived them as less credible or lacking depth—highlighting important design considerations for automated journalism interfaces. [link]

  • Designing an Algorithm-Driven Text Generation System for Personalized and Interactive News Reading (2019), D. Kim, and J. Lee, International Journal of Human-Computer Interaction
    The PINGS system generates personalized, interactive sports news (specifically baseball) based on statistical importance and direct UI manipulation; evaluation found it produces more engaging and pleasant content than traditional news articles. [link]

  • Enhancing Auto-Generated Baseball Highlights via Win Probability and Bias Injection Method (2024), K. Park, H. Lim, J. Lee, and B. Suh, CHI 2024
    Leveraging Win Probability Added (WPA) and bias manipulation, the authors show that WPA-based automated baseball highlights are rated more favorably than existing AI-generated ones; perceived bias and game outcome significantly influence user perception. [link]

Social Computing

Social Computing is a field that examines social interaction and communication in online environments, analyzing how people connect, collaborate, and share information through social media, online communities, and collaborative tools. Our research explores the interplay between technology and social behavior, from understanding public discourse to designing systems that support collective engagement. Recent projects include using natural language processing to compare perceptions of the metaverse in news media and academic research, analyzing media coverage of “social distancing” during the second wave of COVID-19 through deep learning–based news selection, and developing SportLight, a statistically principled crowdsourcing method for sports highlight selection. Through these studies, we aim to better understand how digital platforms shape social interaction and to design tools that foster informed, meaningful, and inclusive online communication.

Research Highlights
  • Analysis of News Coverage on “Social Distancing” during the Second Wave of COVID-19 Using Deep Learning (2021) , G. Lee, and J. Lee, Korean Journal of Journalism and Communication
    This study analyzed over 14,000 news articles across five pandemic phases using frequency analysis, keyword networks, and Top2Vec embedding; it found that “social distancing” dominated featured news more than “COVID-19,” with featured coverage skewed toward political/economic sections in early phases, while non-featured news showed broader topics and sections, especially in later phases. [link]

  • SportLight: Statistically Principled Crowdsourcing Method for Sports Highlight Selection (2022) , J. Jung, S. Ha, W. Son, J. Lee, and J. H. Won, Journal of the Korean Statistical Society
    Introducing SportLight—a method that combines multiple hypothesis testing and ℓ₁-trend filtering (fused lasso)—the authors show that their crowdsourcing approach, applied to 29 baseball games, reduces false alarms and produces highlight selections closer to expert choices than traditional peak-finding methods. [link]