Design Principles of Game AI Assistant (GAIA) for Players with Disabilities: Accessibility Needs and Ethical ConcernsVideo games offer immense potential for identity formation and social participation for players with disabilities; nevertheless, fragmented accessibility settings and assistive tools often impose cycles of uncompensated accessibility labor. We explore the Game AI Assistant (GAIA) concept, designed as a framework for personalized support that should adapt to both game context and inferred needs, while preserving autonomy and fair play. We conducted ethics-approved, semi-structured interviews with seven hardcore players with disabilities. Participants were shown a prototype in the context of a high-concentration fighting game to elicit their perceptions and needs. Thematic analysis revealed a core conflict around the psychological state of flow: the prototype was valued for reducing accessibility labor during low-focus preparatory phases but was rejected as a disruptive intrusion during high-concentration moments. While participants appreciated guidance on accessibility resources, they criticized the prototype's limited customization and verbose responses. Based on these findings, we propose two design principles. The first, Dual Context Adaptation, mandates that an assistant must dynamically adapt its behavior to protect the player's state of flow. The second, an Ethical Framework for Agency and Accomplishment, is grounded in Self-Determination Theory, so the assistant protects the player's fundamental needs for autonomy and competence.2026JCJihun Chae et al.Korea Advanced Institute of Science and TechnologyGame AccessibilityMotor Impairment Assistive Input TechnologiesInclusive DesignIUI
In-Situ Adaptive Interfaces for Online Browsing: Design Dimensions for Intent-Responsive Automation and User ControlOnline browsing often requires balancing open-ended exploration with focused comparison, yet most interfaces remain static regardless of user focus. Adaptive interfaces offer a way to better align interface presentation with browsing needs, but how such adaptations should be triggered, designed, and controlled in practice remains unclear. We investigate adaptive browsing interfaces that modify information hierarchy, information granularity, and session-based ordering in response to inferred browsing intent. Drawing from a framework of five adaptation dimensions, we implemented ReLay, a browser-based probe that applies lightweight, in-situ adaptations automatically while allowing user overrides. In a two-phase study (n = 10), participants welcomed adaptive changes when they were transparent, consistent, and easily reversible. Rather than treating control as error correction, users used it as a means of calibration—testing and tuning automation before accepting it. These findings illustrate how intent-responsive, controllable adaptation can support browsing without diminishing user agency. Our work contributes (1) a conceptual framework for adaptive interface behaviors in online browsing, (2) an instantiation of the framework that selectively operationalizes three dimensions, and (3) empirical insights into user acceptance, control, and design implications for adaptive browsing systems.2026EKEunhye Kim et al.KAISTAI-Assisted Decision-Making & AutomationRecommender System UXInformation Filtering & PersonalizationIUI
ChoiceMates: Supporting Unfamiliar Online Decision-Making with Multi-Agent Conversational InteractionsFrom purchasing a gift to deciding on a hobby, unfamiliar decisions---decisions without domain knowledge and experience---are frequent and significant. The complexity and uncertainty of such decisions demand unique approaches to information seeking, understanding, and decision-making. Our formative study highlights that in the current workflow, users want to start by discovering broad and relevant domain information evenly and simultaneously, quickly address emerging inquiries, and gain personalized standards to assess information found. We present ChoiceMates, an interactive multi-agent system designed to address these needs by enabling users to engage with a dynamic set of LLM agents each presenting a unique experience in the domain. Unlike existing multi-agent systems that automate tasks with agents, the user orchestrates agents to assist their decision-making process in each turn, through chatting with all agents, with a tagged subset of agents, or calling in new agents into the space. By comparing ChoiceMates with a web search condition and a multi-agent framework (n=12), we show that ChoiceMates enables a more confident, satisfactory decision-making with better situation understanding than web search, and higher decision quality than a commercial multi-agent framework. We further illustrate how participants utilized ChoiceMates to make unfamiliar decisions, providing insights into designing a more controllable and collaborative multi-agent system.2026JPJeongeon Park et al.University of California San DiegoHuman-LLM CollaborationAI-Assisted Decision-Making & AutomationConversational ChatbotsIUI
WatchHand: Enabling Continuous Hand Pose Tracking On Off-the-Shelf SmartwatchesTracking hand poses on wrist-wearables enables rich, expressive interactions, yet remains unavailable on commercial smartwatches, as prior implementations rely on external sensors or custom hardware, limiting their real-world applicability. To address this, we present WatchHand, the first continuous 3D hand pose tracking system implemented on off-the-shelf smartwatches using only their built-in speaker and microphone. WatchHand emits inaudible frequency-modulated continuous waves and captures their reflections from the hand. These acoustic signals are processed by a deep-learning model that estimates 3D hand poses for 20 finger joints. We evaluate WatchHand across diverse real-world conditions---multiple smartwatch models, wearing-hands, body postures, noise conditions, pose-variation protocols---and achieve a mean per-joint position error of 7.87 mm in cross-session tests with device remounting. Although performance drops for unseen users or gestures, the model adapts effectively with lightweight fine-tuning on small amounts of data. Overall, WatchHand lowers the barrier to smartwatch-based hand tracking by eliminating additional hardware while enabling robust, always-available interactions on millions of existing devices.2026JKJiwan Kim et al.KAISTHand Gesture RecognitionSmartwatches & Fitness BandsContext-Aware ComputingCHI
Auditorily Embodied Conversational Agents: Effects of Spatialization and Situated Audio Cues on Presence and Social PerceptionEmbodiment can enhance conversational agents, such as increasing their perceived presence. This is typically achieved through visual representations of a virtual body; however, visual modalities are not always available, such as when users interact with agents using headphones or display-less glasses. In this work, we explore auditory embodiment. By introducing auditory cues of bodily presence — through spatially localized voice and situated Foley audio from environmental interactions — we investigate how audio alone can convey embodiment and influence perceptions of a conversational agent. We conducted a 2 (spatialization: monaural vs. spatialized) × 2 (Foley: none vs. Foley) within-subjects study, where participants (n=24) engaged in conversations with agents. Our results show that spatialization and Foley increase co-presence, but reduce users’ perceptions of the agent’s attention and other social attributes.2026YCYi Fei Cheng et al.Carnegie Mellon UniversityAffective Human-Computer DialogueSpatial Audio & 3D SoundAffective Feedback & Emotion Regulation InterfacesCHI
AI meets Mathematics Education: Supporting Instructors in Large Mathematics Classes with Context-Aware AILarge-enrollment university courses face persistent challenges in providing timely and scalable instructional support. While generative AI holds promise, its effective use depends on reliability and pedagogical alignment. We present a human-centered case study of AI-assisted support in a Calculus I course, implemented in close collaboration with the course instructor. We developed a system to answer students’ questions on a discussion forum, fine-tuning a lightweight language model on 2,588 historical student–instructor interactions. The model achieved 75.3% accuracy on a benchmark of 150 representative questions annotated by five instructors, and in 36% of cases, its responses were rated equal to or better than instructor answers. Post-deployment student survey (N = 105) indicated that students valued the alignment of the responses with the course materials and their immediate availability, while still relying on the instructor verification for trust. We highlight the importance of hybrid human–AI workflows for safe and effective course support.2026JBJérémy Valentin Barghorn et al.EPFLHuman-LLM CollaborationIntelligent Tutoring Systems & Learning AnalyticsCHI
What Are You Really Asking For? A Comparative 5W1H Analysis of Learner Questioning in CPR Training with IVAs in Screen-based and Augmented Reality EnvironmentsQuestion-asking is one of the key indicators of cognitive engagement. However, understanding how the distinct psychological affordances of presentation media shape learners' spoken inquiries with embodied Intelligent Virtual Agents (IVAs) remains limited. To systematically examine this process, we propose a 5W1H-based framework for analyzing learner questions. Using this framework, we conducted a user study comparing an Augmented Reality-based IVA (AR-IVA) deployed in the physical environment with a screen-based IVA (Video-IVA) during cardiopulmonary resuscitation (CPR) instruction. Results showed that the AR-IVA elicited higher spatial and social presence and promoted more frequent and longer questions focused on clarification and understanding. In contrast, the Video-IVA encouraged questions regarding procedural refinement. Presence acted as a selective filter, shaping the timing and topic of questions rather than as a universal mediator. These effects were significantly moderated by learners’ motivational and strategic characteristics toward learning. Based on these findings, we propose design implications for IVA-supported learning systems.2026HPHYERIM PARK et al.KAISTVR Medical Training & RehabilitationSocial & Collaborative VRIntelligent Tutoring Systems & Learning AnalyticsCHI
GazeZoom: Exploration of Gaze-Assisted Multimodal Techniques for Panning and ZoomingZooming and panning are fundamental input actions for exploring complex 2D and 3D scenes and data such as images, maps, and designs. Multi-touch zoom/pan interactions have been proven effective on mobile devices, and have been directly ported to HMDs, where they are typically accomplished by analogous but relatively large-scale movements of both hands. We argue that such motions are inefficient and induce fatigue and explore how the eye-tracking features of HMDs can be leveraged to achieve improvements. We evaluated three interaction techniques that combine gaze with two-handed, one-handed, and head-based input in a study (N=24) that contrasts them against a baseline two-handed technique. The results indicate that gaze-assisted two- and one-handed techniques outperform the baseline (17%-36% faster), while our head-based technique achieves similar performance to the Baseline but leaves the hands free for other tasks. We further developed a VR application demonstrating these techniques and validating their practical applicability.2026YLYilong Lin et al.Southern University of Science and TechnologyEye Tracking & Gaze InteractionSocial & Collaborative VRImmersion & Presence ResearchCHI
ClearFairy: Capturing Creative Workflows through Decision Structuring, In-Situ Questioning, and Rationale InferenceCapturing professionals’ decision-making in creative workflows (e.g., UI/UX) is essential for reflection, collaboration, and knowledge sharing, yet existing methods often leave rationales incomplete and implicit decisions hidden. To address this, we present the CLEAR approach, which structures reasoning into cognitive decision steps—linked units of actions, artifacts, and explanations, making decisions traceable with generative AI. Building on CLEAR, we introduce ClearFairy, a think-aloud AI assistant for UI design that detects weak explanations, asks lightweight clarifying questions, and infers missing rationales. In a study with twelve professionals, 85% of ClearFairy’s inferred rationales were accepted (as-is or with revisions). Notably, the system increased "strong explanations"'—rationales providing sufficient causal reasoning—from 14% to 83% without adding cognitive demand. Furthermore, exploratory applications demonstrate that captured steps can enhance generative AI agents in Figma, yielding predictions better aligned with professionals and producing coherent outcomes. We release a dataset of 417 decision steps to support future research.2026KSKihoon Son et al.KAISTHuman-LLM CollaborationCreative Collaboration & Feedback Systems360° Video & Panoramic ContentCHI
“Don’t Look, But I Know You Do”: Norms and Observer Effects in Shared LLM AccountsAccount sharing is common in subscription services and is now extending to generative AI platforms, which are still primarily designed for individual use. Sharing often requires workarounds that create new tensions. This study examines how LLM subscriptions are shared and the norms that develop. We combined a survey of 245 users with interviews of 36 participants to understand both patterns and lived experiences. Our analysis identified four types of account sharing, organized along two dimensions: whether the owner uses the account and whether subscription costs are shared. Within these types, we examined how norms were formed and how their fragility, especially privacy, became evident in practice. Users, fully aware of this, subtly adjusted their behavior, which we interpret through the lens of the observer effect. We frame LLM account sharing as a social practice of appropriation and outline design implications to adapt single-user platforms to multi-user realities.2026JSJi Eun Song et al.Seoul National UniversityHuman-LLM CollaborationPrivacy by Design & User ControlPrivacy Perception & Decision-MakingCHI
Exploring Data-Driven Approaches to Stress Management: A Systematic Review of Stress Tracking, Intervention, and System Evaluation MethodsAdvances in ubiquitous and wearable sensing and HCI research have made stress monitoring increasingly accessible, enabling the development of personalized stress management technologies. Yet, stress is a subjective and contextual experience, making effective intervention design challenging. Prior studies often isolate stress detection or intervention, without providing an integrated view of how these components connect and are evaluated in real-world use. To address this gap, we conducted a systematic review of 2,152 papers and selected 52 empirical studies where stress tracking informed interventions. Using a framework based on three stress constructs (subjective stress, psycho-physiological stress, and exposure stress), we analyzed how definitions of stress shape detection indicators, intervention design and timing, and evaluation methods. We show that stress conceptualization strongly influences system design, and we propose a conceptual framework linking detection, intervention, and evaluation to guide future user-centered stress management technologies.2026YKYoungji Koh et al.KAISTSleep & Stress MonitoringHealth Self-TrackingBehavior Change & Reflection TechnologyCHI
TingleTouch: Touch Guidance through Electrical Stimulation in Resistance TrainingIn resistance training, trainers employ touch guidance to help trainees control posture and activate muscles. Haptic feedback can extend this support to solitary workouts, but translating the nuances of touch into effective haptic patterns remains challenging. In this paper, we categorize the instructional messages conveyed through trainers' touch guidance and design electrical stimulation patterns to replicate them. A preliminary study with six trainers and six trainees identified six core messages underlying touch guidance. We then designed electrical stimulation patterns for each message and refined them with two sports scientists and a UX designer, ensuring usability and grounding. Finally, sixteen gymgoers evaluated these patterns in a controlled exercise task. Participants reliably distinguished the feedback and used the instructed muscles accordingly, achieving accuracies of 97.14% and 99.22% across two sessions, cross-checked with EMG and pose estimation. These findings demonstrate that the proposed electrical stimulation feedback is intuitive and learnable.2026DKDong-Uk Kim et al.Chung-Ang UniversityElectrical Muscle Stimulation (EMS)Fitness Tracking & Physical Activity MonitoringBehavior Change & Reflection TechnologyCHI
TreeB612: Exploring Interactions for Distant Human–Nature EngagementOpportunities for people to connect with everyday nature have diminished. In response, research efforts have introduced both direct and remote approaches to improve human–nature connections. However, most work has relied heavily on visual and auditory media as the main channels for remote human–nature interfaces. In this study, we investigate what experiential factors and forms of human–nature connection emerge from an interface without screens or audio. To inform design, we surveyed 55 respondents to understand expectations for remote nature engagement. Guided by these insights, we designed TreeB612, enabling one-on-one engagement with a distant living entity—a single tree. We conducted a 7-day deployment study with 10 participants, using diaries to explore how the interface shaped everyday engagement. Our findings suggest that the interface fostered coexistence and care, evoked the chosen tree through imagination and moments of respite, and encouraged subtle shifts in how people related to nature, advancing approaches to nature connection.2026GKGurim Kim et al.KAISTHuman-Nature Relationships (More-than-Human Design)Tangible User Interface DesignPhysical-Digital Hybrid InteractionCHI
Building Human–Multi-Agent Teams for Creative WorksTeam-based collaboration is a cornerstone of modern creative work. Recent advances in generative AI open possibilities for humans to collaborate with multiple AI agents in distinct roles to address complex creative workflows. Yet, how to form Human–Multi-Agent Teams (HMATs) is underexplored, especially given that inter-agent interactions increase complexity and the risk of unexpected behaviors. In this exploratory study, we aim to understand how to form HMATs for creative work using CrafTeam, a technology probe that allows users to form and collaborate with their teams. We conducted a study with 12 design practitioners, in which participants iterated through a three-step cycle: forming HMATs, ideating with their teams, and reflecting on their teams' ideation. Our findings reveal that while participants initially attempted autonomous team operations, they ultimately adopted team formations in which they directly orchestrated agents. We discuss design considerations for HMAT formation that humans can effectively orchestrate multiple agents.2026HLHyunseung Lim et al.KAISTGenerative AI (Text, Image, Music, Video)Creative Collaboration & Feedback SystemsCHI
Evalet: Evaluating Large Language Models through Functional FragmentationPractitioners increasingly rely on Large Language Models (LLMs) to evaluate generative AI outputs through "LLM-as-a-Judge" approaches. However, these methods produce holistic scores that obscure which specific elements influenced the assessments. We propose functional fragmentation, a method that dissects each output into key fragments and interprets the rhetoric functions that each fragment serves relative to evaluation criteria—surfacing the elements of interest and revealing how they fulfill or hinder user goals. We instantiate this approach in Evalet, an interactive system that visualizes fragment-level functions across many outputs to support inspection, rating, and comparison of evaluations. A user study (N=10) found that, while practitioners struggled to validate holistic scores, our approach helped them identify 48% more evaluation misalignments. This helped them calibrate trust in LLM evaluations and rely on them to find more actionable issues in model outputs. Our work shifts LLM evaluation from quantitative scores toward qualitative, fine-grained analysis of model behavior.2026TKTae Soo Kim et al.KAISTHuman-LLM CollaborationExplainable AI (XAI)User Research Methods (Interviews, Surveys, Observation)CHI
Fostering Collective Discourse: A Distributed Role-Based Approach to Online News CommentingCurrent news commenting systems are designed based on implicitly individualistic assumptions, where discussion is the result of a series of disconnected opinions. This often results in fragmented and polarized conversations that fail to represent the spectrum of public discourse. In this work, we develop a news commenting system where users take on distributed roles to collaboratively structure the comments to encourage a connected, balanced discussion space. Through a within-subject, mixed-methods evaluation (N=38), we find that the system supported three stages of participation: understanding issues, collaboratively structuring comments, and building a discussion. With our system, users' comments displayed more balanced perspectives and a more emotionally neutral argumentation. Simultaneously, we observed reduced argument strength compared to a traditional commenting system, indicating a trade-off between inclusivity and depth. We conclude with design considerations and trade-offs for introducing distributed roles in news commenting system design.2026YHYoojin Hong et al.KAISTContent Moderation & Platform GovernanceSocial Platform Design & User BehaviorCHI
Constructing Everyday Well-Being: Insights from God-Saeng (God生) for Personal InformaticsWhile Personal Informatics (PI) systems support behavior change, everyday well-being involves more than achieving individual target behaviors. It is shaped by cultural narratives that give actions meaning. In South Korea, the God-Saeng (God生) phenomenon—encompassing disciplined, collective, and publicly documented self-improvement practices—offers a lens into how well-being is negotiated in daily life. We conducted a 10-day probe (N=24) with bite-sized missions to examine how young adults engaged in God-Saeng. Participants relied on planning practices, accountability infrastructures, and datafication to stabilize themselves, yet these same routines also intensified pressures toward self-monitoring and performance. They navigated tensions between consistency and flexibility, authenticity and visibility, and productivity and broader values such as relationships, and reinterpreted ordinary activities through sociocultural contexts. These insights suggest design opportunities for PI systems that move beyond tracking, toward digital instruments that help users negotiate tensions, make meaning, and reflexively understand how technologies participate in their culturally and existentially situated well-being.2026ISInhwa Song et al.Princeton UniversityBehavior Change & Reflection TechnologyData-Driven Personal Decision-MakingInclusive DesignCHI
When Plants Play: Rethinking Plant Materiality in Digital GamesPlants are rarely positioned as active participants in digital games, serving as decorative elements or passive sensors. We present Plant.play(), a plant–digital game system that positions a living plant as the sole player in a pet-simulation game. Using bioelectrical signals, environmental data, and circadian rhythms, the plant autonomously performs caregiving actions while humans engage as observers. A workshop with five game experts informed the system's design, which was then implemented and deployed in a four-day exhibition. Observational fieldwork and interviews with twelve visitors revealed how people initially sought control, then gradually shifted toward interpreting the plant's slow, unpredictable, and impartial behaviors as meaningful play. Participants formed emotional connections with both the plant and the virtual pet, extending these reflections to their relationships with nonhuman beings. Our findings contribute empirical insights into interpretive engagement with nonhuman actors and offer design considerations for future plant–digital game systems that embrace materiality, perceived agency, and more-than-human perspectives.2026YLYoonji Lee et al.Korea Advanced Institute of Science and Technology (KAIST)Game UX & Player BehaviorShape-Changing Interfaces & Soft Robotic MaterialsHuman-Nature Relationships (More-than-Human Design)CHI
Gaze and Speech in Multimodal Human-Computer Interaction: A Scoping ReviewMultimodal interaction has long promised to make interfaces more intuitive and effective by combining complementary inputs. Among these, gaze and speech form a compelling pairing: gaze provides rapid spatial grounding, while speech conveys rich semantic information. Together, they offer rich cues for understanding user behaviour and intent. Yet despite decades of exploration, the research remains fragmented, making this synthesis timely as these inputs mature and are integrated into consumer-ready devices. This scoping review examined 103 studies published between 1991 and 2025, organised into \emph{explicit}, where users intentionally provide gaze and speech, and \emph{implicit}, where systems leverage users' natural behaviours to support interaction. Across both, we identified recurring ways for combining gaze and speech to resolve ambiguity, ground references, and support adaptivity. We contribute a synthesis of research on their combined use while highlighting challenges of temporal alignment, fusion and privacy, offering guidance for future research toward richer multimodal human-computer interaction.2026AKAnam Ahmad Khan et al.KAISTEye Tracking & Gaze InteractionVoice User Interface (VUI) DesignAffective Human-Computer DialogueCHI
ComVi: Context-Aware Optimized Comment Display in Video PlaybackOn general video-sharing platforms like YouTube, comments are displayed independently of video playback. As viewers often read comments while watching a video, they may encounter ones referring to moments unrelated to the current scene, which can reveal spoilers and disrupt immersion. To address this problem, we present ComVi, a novel system that displays comments at contextually relevant moments, enabling viewers to see time-synchronized comments and video content together. We first map all comments to relevant video timestamps by computing audio-visual correlation, then construct the comment sequence through an optimization that considers temporal relevance, popularity (number of likes), and display duration for comfortable reading. In a user study, ComVi provided a significantly more engaging experience than conventional video interfaces (i.e., YouTube and Danmaku), with 71.9% of participants selecting ComVi as their most preferred interface.2026MKMinsun Kim et al.KAISTSocial & Collaborative VRImmersion & Presence ResearchLive Streaming & Content CreatorsCHI