Toward Scalable and Responsible Integration of Course-Specific AI Tutors: Instructor Experiences with a Campus-Wide Platform Despite rapid investment in generative AI across higher education, how instructors create, evaluate, and implement course-specific AI tutors remain empirically underexplored, highlighting critical tensions between institutional adoption and instructional practices. Drawing on interviews with 20 instructors, teaching assistants, and instructional designers at a large U.S. research university, we examine how participants engaged with a university-wide platform for creating course-specific AI tutors. Our findings reveal how instructors’ epistemic beliefs and pedagogical orientations shaped their perceptions of appropriate and inappropriate AI uses, as well as how instructional challenges motivated tutor creation across disciplines, class sizes, and course levels. We also identified three key patterns in instructor evaluation of course-specific AI tutors, along with the pedagogical, technical, and ethical implementation challenges they faced. We contribute timely insights to inform research, platform development, and institutional policy toward the responsible and scalable integration of course-specific AI tutors in higher education.2026EKEunhye Grace Ko et al.University of Texas at AustinHuman-LLM CollaborationAI-Assisted Decision-Making & AutomationAI Ethics, Fairness & AccountabilityCHI
"I Can Be Anything!" Bridging Today and the Future through Generative AI-driven Self-represented Career Imagination for ChildrenChildren’s career imagination of their future selves is critical for motivation and identity formation, particularly by fostering continuity between present and future. However, current interventions often lead children—whose developmental stage limits future-oriented thinking—to view their future not as something connected to themselves, but as a detached job title or label. We present FutureMe, a generative AI–driven system that helps children imagine self-represented futures. Through a step-by-step process of taking a photo, exploring career options, imagining actions, and building story strips, children see themselves embedded in AI-generated narratives. The activity concludes with writing letters to their future selves. In a study with 17 children, through FutureMe, children revealed hidden aspirations, integrated past and present experiences into imagined futures, and redefined externally prescribed “good careers” as personally meaningful choices. We discuss implications for designing AI tools that empower children’s agency in imagining futures.2026SLSunok Lee et al.Sogang UniversityGenerative AI (Text, Image, Music, Video)Children's AI Literacy & Data LiteracyProgramming Education & Computational ThinkingCHI
MUST: Smartwatch-based Multimodal Framework for Predicting Driver State and Takeover PerformanceEnsuring timely takeover in conditionally autonomous vehicles presents a significant challenge, especially when drivers are distracted by non-driving-related tasks or are in suboptimal emotional states. Existing driver monitoring systems struggle with a trade-off between practicality and reliability. Physiological sensors are intrusive, vision-based methods are sensitive to occlusions and variable lighting, and current multimodal learning approaches often rely on simple fusion strategies that fail to reconcile heterogeneous data. We introduce MUST (Multimodal Unified Smartwatch-based Takeover), a framework that predicts driver state and takeover performance using unobtrusive smartwatch signals. MUST employs an asymmetric causal fusion mechanism to model the interplay between driver behavior and emotion. The performance of the architecture was validated in diverse simulator environments reflecting real-world driving conditions, demonstrating robust driver state estimation and takeover prediction. This work establishes the smartwatch as a practical tool for adaptive takeover support, enabling reliable readiness assessment without intrusive hardware or fragile vision systems.2026SSSeokyong Sheem et al.Korea UniversityAutomated Driving Interface & Takeover DesignIn-Vehicle Haptic, Audio & Multimodal FeedbackSmartwatches & Fitness BandsCHI
A Shared Look: Detecting Deepfakes with Inter-Subject Neural SynchronyThe rapid evolution of generative AI presents a significant challenge for Deepfake detection. While most research focuses on face-swapping, the emerging threat of "image-to-video" (I2V) forgeries is harder to detect and poses a greater risk. Traditional computer vision detectors rely on transient digital artifacts, which often lack interpretability and robustness against the new generation techniques. This study introduces a neuro-cognitive method, using dyadic electroencephalogram (EEG) to decode the human perception of authenticity. We recorded inter-brain synchrony via EEG hyperscanning from 15 participant pairs as they viewed a balanced set of authentic and AI-generated videos. Results showed that these shared neural response can classify video authenticity with an accuracy of up to 89.23% using our proposed Hyper-FusionNet. In addition, the biomarkers exhibited distinct patterns for different emotional valences, highlighting their versatility. These findings highlight the potential of inter-brain synchrony for detecting emerging deepfakes, offering a new perspective for enhancing user trust and digital literacy.2026SHShiang Hu et al.Anhui UniversityDeepfake & Synthetic Media DetectionEmotion Recognition & DetectionAffective Feedback & Emotion Regulation InterfacesCHI
Debugging Defective Visualizations: Empirical Insights Informing a Human-AI Co‑Debugging SystemVisualization authoring is an iterative process requiring users to adjust parameters to achieve desired aesthetics. Due to its complexity, users often create defective visualizations and struggle to fix them. Many seek help on forums (e.g., Stack Overflow), while others turn to AI, yet little is known about the strengths and limitations of these approaches, or how they can be effectively combined. We analyze Vega-Lite debugging cases from Stack Overflow, categorizing question types by askers, evaluating human responses, and assessing AI performance. Guided by these findings, we design a human-AI co-debugging system that combines LLM-generated suggestions with forum knowledge. We evaluated this system in a user study on 36 unresolved problems, comparing it with forum answers and LLM baselines. Our results show that while forum contributors provide accurate but slow solutions and LLMs offer immediate but sometimes misaligned guidance, the hybrid system resolves 86\% of cases, higher than either alone.2026SSShuyu Shen et al.Hong Kong University of Science and Technology (Guangzhou)Interactive Data VisualizationHuman-LLM CollaborationAI-Assisted Decision-Making & AutomationCHI
"Can LLMs Persuade Humans with Deception?": From a Deceptive Strategy Taxonomy to a Large-Scale Empirical StudyBeyond hallucinations, Large Language Models (LLMs) can craft deceptive arguments that erode users' critical thinking, posing a significant yet underexamined societal risk. To address this gap, we develop a taxonomy of eight deceptive persuasion strategies by integrating top-down rhetorical theory with a bottom-up analysis of 3,360 AI-generated messages by four LLM families and examining their effects on user perceptions. Through a large-scale user study (N=602) complemented by a think-aloud protocol, we found that participants were vulnerable to \textit{Information Manipulation} and \textit{Uncertainty Exploitation}, especially when a message contradicted their prior beliefs. Vulnerability was significantly higher for participants with low cognitive reflection, low topic knowledge, and low topic involvement. Qualitative analyses further revealed that participants were persuaded by the plausibility of an overall narrative even when they distrust specific details, interpreting deceptive outputs as logically framed information that broadens perspective. We discuss critical implications of these findings for the design of trustworthy AI systems, adaptive user interfaces, and targeted literacy education.2026HYHaein Yeo et al.Hanyang UniversityAI Ethics, Fairness & AccountabilityExplainable AI (XAI)Privacy by Design & User ControlCHI
Educator Perceptions of XRAuthor: An Accessible Tool for Authoring Learning Content with Different Immersion LevelsThe promise of Extended Reality (XR) in education is significant but one size does not fit all learning contexts and student preferences. Varied content with different immersion levels is hence beneficial, but creating XR content remains daunting for educators using conventional tools. This paper introduces XRAuthor, a web-based authoring tool designed to empower educators to create varying immersive learning content - ranging from conventional video to interactive animations and full-fledged VR - all from a single authoring experience with a webcam. Through online one-to-one workshops with 14 educators, we found strong endorsement for the new authoring workflow enabled by XRAuthor. Participants also found that the varied interactive exercises automatically generated by the tool aligned well with effective pedagogical practices. High ease of use and efficiency were identified as crucial attributes of XRAuthor. The design knowledge facilitated by XRAuthor underscores the potential of such tool designs to democratize XR content creation for learning.2025SSSongjia Shen et al.Singapore Institute of Technology, Centre for ImmersificationMixed Reality WorkspacesOnline Learning & MOOC PlatformsCHI
Exploring the Impact of Avatar Representations in AI Chatbot Tutors on Learning ExperiencesDespite the growing prominence of Artificial Intelligence (AI) chatbots used in education, there remains a significant gap in our understanding of how interface design elements, particularly avatar representations, influence learning experiences. This paper explores the impact of different AI chatbot avatar representations on students' learning experiences through a mixed-methods within-subjects study, where participants interacted with three distinct types of AI chatbot interfaces with a common large language model (LLM) over a 14-week university course. Our findings reveal that preferences vary according to factors such as learning habits and learning activities. Avatar design also exhibits affordances for specific prompting behaviors, while the perceived human touch influenced learning experiences in nuanced ways. Additionally, real-world relationships with the individuals behind deepfakes influence these experiences. These insights suggest that the thoughtful integration of diverse avatar representations in AI chatbot systems for different learners and settings can greatly enhance learning experiences.2025CTChek Tien Tan et al.Singapore Institute of Technology, Centre for ImmersificationAgent Personality & AnthropomorphismHuman-LLM CollaborationCHI
HearFire: Indoor Fire Detection via Inaudible Acoustic Sensing"Indoor conflagration causes a large number of casualties and property losses worldwide every year. Yet existing indoor fire detection systems either suffer from short sensing range (e.g., ≤ 0.5m using a thermometer), susceptible to interferences (e.g., smoke detector) or high computational and deployment overhead (e.g., cameras, Wi-Fi). This paper proposes HearFire, a cost-effective, easy-to-use and timely room-scale fire detection system via acoustic sensing. HearFire consists of a collocated commodity speaker and microphone pair, which remotely senses fire by emitting inaudible sound waves. Unlike existing works that use signal reflection effect to fulfill acoustic sensing tasks, HearFire leverages sound absorption and sound speed variations to sense the fire due to unique physical properties of flame. Through a deep analysis of sound transmission, HearFire effectively achieves room-scale sensing by correlating the relationship between the transmission signal length and sensing distance. The transmission frame is carefully selected to expand sensing range and balance a series of practical factors that impact the system's performance. We further design a simple yet effective approach to remove the environmental interference caused by signal reflection by conducting a deep investigation into channel differences between sound reflection and sound absorption. Specifically, sound reflection results in a much more stable pattern in terms of signal energy than sound absorption, which can be exploited to differentiate the channel measurements caused by fire from other interferences. Extensive experiments demonstrate that HireFire enables a maximum 7m sensing range and achieves timely fire detection in indoor environments with up to 99.2% accuracy under different experiment configurations. https://doi.org/10.1145/3569500"2023ZWZheng Wang et al.Context-Aware ComputingUbiquitous ComputingUbiComp
"We Gather Together, We Collaborate Together": Exploring the Challenges and Strategies of Chinese Lesbian and Bisexual Women's Online Communities on WeiboIn China, lesbian and bisexual women face intense stigma and difficulties developing relationships with each other. Although prior research has shown that online communities help LGBT people connect and exchange social support, few studies have explored the challenges Chinese lesbian and bisexual women face when initiating, growing, and sustaining such communities, in an atmosphere of platform censorship of LGBT-related content and intense discrimination from non-LGBT people. To address this gap, we interviewed 40 Weibo users in China, four bloggers and 36 followers of their blogs, who self-identified as lesbian or bisexual women. We found that a key technique these bloggers used to initiate their online communities was helping followers publish posts seeking support, sharing personal experiences, and seeking offline relationships. Then, their followers built relationships with bloggers by journaling their daily experiences as lesbian or bisexual women via private-messaging channels. As the communities’ members grew more attached to them, bloggers and their followers began to work together to protect themselves from external threats, including Weibo’s censorship and non-LGBT+ infiltrators’ harassment. However, such attachment to the communities sometimes might lead to conflicts within them, which in turn prompted many members to leave, raising questions about the communities’ long-term prospects. Our findings foreground important design considerations for those seeking to help lesbian and bisexual women in China and other discriminatory environments to develop safe online communities.2022YCYichao Cui et al.Online Communities & Inclusivity; Online Communities & InclusivityCSCW
Understanding User Experiences Across VR Walking-in-place Locomotion MethodsNavigating large-scale virtual spaces is a major challenge in Virtual Reality (VR) applications due to real-world spatial limitations. Walking-in-place (WIP) locomotion solutions may provide a natural approach for VR use cases that require locomotion to share similar qualities with walking in real-life. However, there is limited knowledge on the range of experiences across common WIP methods to inform the design of usable WIP solutions using consumer-accessible components. This paper contributes to this knowledge via a user study with 40 participants that experienced several easy-to-setup WIP methods in a VR commuting simulation. A nuanced understanding of cybersickness and exertion relationships and walking affordances based on different tracker setups were among the findings derived from a corroborated analysis of think-aloud, interview, and observational data, supplemented with self-reports of VR sickness, presence and flow. Practical design insights were then constructed along the dimensions of cybersickness, affordances, space and user interfaces.2022CTChek Tien Tan et al.Singapore Institute of TechnologyFull-Body Interaction & Embodied InputSocial & Collaborative VRImmersion & Presence ResearchCHI
Attending to Slowness and Temporality with Olly and Slow Game: A Design Inquiry Into Supporting Longer-Term Relations With Everyday Computational ObjectsSlowness has emerged as a rich lens to frame HCI investigations into supporting longer-term human-technology relations. Yet, there is a need to further address how we design for slowness on conceptual and practical levels. Drawing on the concepts of unawareness, intersections, and ensembles, we contribute an investigation into designing for slowness and temporality grounded in design practice through two cases: Olly and Slow Game. We designed these artifacts over two and a half years with careful attention to how the set of concepts influenced key design decisions in terms of their form, materials, and computational qualities. Our designer-researcher approach revealed that, when put into practice, the concepts helped generatively grapple with slowness and temporality, but are in need of further development to be mobilized for design. We critically reflect on insights emerging across our practice-based research to reflexively refine the concepts and better support future HCI research and practice.2018WOWilliam Odom et al.Simon Fraser UniversityTechnology Ethics & Critical HCIDesign FictionCHI