The Invisible Mentor: Inferring User Actions from Screen Recordings to Recommend Better WorkflowsUsers of feature-rich tools like Excel often miss more efficient workflows, repeating tedious steps and making avoidable errors. Current approaches to helping them require either manual prompting, which is effortful for users, or automated logging, which is limiting for developers. We present InvisibleMentor, a system inspired by over-the-shoulder learning: it observes what users do, then shows them how to do it better. To do this, InvisibleMentor analyzes screen recordings with a vision-language model to reconstruct actions and context, then uses a large language model to generate vision-grounded task reflection, structured suggestions grounded in observed behavior. In a user study, participants found InvisibleMentor's suggestions more clear, more relevant, and more useful than those from a prompt-based assistant, demonstrating that AI can do more than automate away work—it can help users master it.2026LYLitao Yan et al.University of PennsylvaniaHuman-LLM CollaborationAI-Assisted Decision-Making & AutomationUser Research Methods (Interviews, Surveys, Observation)CHI
Laughing Through the Struggles: Understanding ADHD Experience and Community Engagement Through Memes and Comments on InstagramWhile public discourse often reduces Attention-Deficit Hyperactivity Disorder (ADHD) to stereotypes that overlook the invisible struggles of those who live with it, ADHD people are increasingly using social media to express their experiences on their own terms. On platforms like Instagram, memes have become a powerful and accessible medium for expressing everyday challenges through humor and relatability. This study analyzed 350 ADHD-related memes and over 28,000 associated comments to explore how ADHD was expressed and engaged with in online spaces, and consulted a neurodevelopmental science and clinical researcher. Findings show that memes depict behavioral inconsistencies, internal conflicts, and societal pressures, while comments reveal strong resonance, personal identification, and peer support, including informal self-diagnosis and shared experiences. By combining meme and comment analyses, this study contributes to digital mental health research by demonstrating how memes serve as an interactional mechanism for neurodivergent storytelling and identity formation and informing future platform design.2026FZFan Zhang et al.Independent ResearcherCognitive Impairment & Neurodiversity (Autism, ADHD, Dyslexia)Social Platform Design & User BehaviorMental Health Apps & Online Support CommunitiesCHI
Voice-Based Chatbots for English Speaking Practice in Multilingual Low-Resource Indian Schools: A Multi-Stakeholder StudySpoken English proficiency is a powerful driver of economic mobility for low-income Indian youth, yet opportunities for spoken practice remain scarce in schools. We investigate the deployment of a voice-based chatbot for English conversation practice across four low-resource schools in Delhi. Through a six-day field study combining observations and interviews, we captured the perspectives of students, teachers, and principals. Findings confirm high demand across all groups, with notable gains in student speaking confidence. Our multi-stakeholder analysis surfaced a tension in long-term adoption vision: students favored open-ended conversational practice, while administrators emphasized curriculum-aligned assessment. We offer design recommendations for voice-enabled chatbots in low-resource multilingual contexts, highlighting the need for more intelligible speech output for non-native learners, one-tap interactions with simplified interfaces, and actionable analytics for educators. Beyond language learning, our findings inform the co-design of future AI-based educational technologies that are socially sustainable within the complex ecosystem of low-resource schools.2026SSSneha Shashidhara et al.Centre for Social and Behaviour Change, Ashoka UniversityIntelligent Voice Assistants (Alexa, Siri, etc.)Multilingual & Cross-Cultural Voice InteractionSpecial Education TechnologyCHI
Patchworking Networks of Support: On the Digital Successes and Challenges of Women- and Minority-Owned Restaurant BusinessesThe restaurant industry has become increasingly reliant on digital technologies for business operations, digital marketing, and promotion, especially amid and after the Covid-19 pandemic. This paper presents the findings of a two-year study exploring how women- and minority-owned restaurants in Chicago and Detroit encountered and overcame digital challenges in their day-to-day operations, across a range of levels of digital skills and literacy. Drawing from semi-structured and impromptu interviews with restaurant owners (n=47) and participant observation, we apply HCI literature on infrastructuring and patchworking to highlight how restaurateurs' experiences often run counter to the assumptions of a "typical" user. Indeed, they often must build and leverage their—offline and online—networks of support to overcome failing infrastructures, both within the restaurant industry and on digital platforms. Concurrently, we emphasize the importance of community building and social infrastructuring to overcome these challenges while also building up alternative networks of resources for their communities, especially considering identity-related inequalities and amid a global moment of crisis.2026MBMatthew Bui et al.University of MichiganDeveloping Countries & HCI for Development (HCI4D)Inclusive DesignCommunity Engagement & Civic TechnologyCHI
BiasViz: A Project-Based, Narrative-Centered Learning Tool for Engaging Middle School Students in Critical Thinking about AI BiasesDeveloping the ability to think critically about AI and interpret its outputs requires an understanding of AI bias, a key skill for both AI users and future developers. While some initiatives have introduced teens to algorithmic bias, few have engaged them in actively identifying and quantifying bias in real-world generative AI systems. This paper presents BiasViz, an interactive tool that leverages project-based and narrative-centered learning to help middle school students (11-14 year old) analyze AI bias in large language models. We conducted a study of 28 students’ interactions with BiasViz to evaluate its efficacy in fostering critical thinking about AI bias. Our findings suggest that BiasViz successfully introduced most students to AI bias, and some used the tool to explore personally relevant biases. We identify opportunities for the tool’s iteration and associated curriculum to promote learning and share insights for designing learning environments that foster youth’s critical thinking about AI.2026HDHasti Darabipourshiraz et al.Northwestern UniversityHuman-LLM CollaborationAI Ethics, Fairness & AccountabilityProgramming Education & Computational ThinkingCHI
Does Personalized Nudging Wear Off? A Longitudinal Study of AI Self-Modeling for Behavioral EngagementSustaining the effectiveness of behavior change technologies remains a key challenge. AI self-modeling, which generates personalized portrayals of one’s ideal self, has shown promise for motivating behavior change, yet prior work largely examines short-term effects. We present one of the first longitudinal evaluations of AI self-modeling in fitness engagement through a two-stage empirical study. A 1-week, three-arm experiment (visual self-modeling (VSM), auditory self-modeling (ASM), Control; N=28) revealed that VSM drove initial performance gains, while ASM showed no significant effects. A subsequent 4-week study (VSM vs. Control; N=31) demonstrated that VSM sustained higher performance levels but exhibited diminishing improvement rates after two weeks. Interviews uncovered a catalyst effect that fostered early motivation through clear, attainable goals, followed by habituation and internalization which stabilized performance. These findings highlight the temporal dynamics of personalized nudging and inform the design of behavior change technologies for long-term engagement.2026QHQing He et al.University of PennsylvaniaBehavior Change & Reflection TechnologyHealth Self-TrackingEmotion-Sensing WearablesCHI
VizCrit: Exploring Strategies for Displaying Computational Feedback in a Visual Design ToolVisual design instructors often provide multi-modal feedback, mixing annotations with text. Prior theory emphasizes the importance of actionable feedback, where “actionability” lies on a spectrum—from surfacing relevant design concepts to suggesting concrete fixes. How might creativity tools implement annotations that support such feedback, and how does the actionability of feedback impact novices’ process-related behaviors, perceptions of creativity, learning of design principles, and overall outcomes? We introduce VizCrit, a system for providing computational feedback that supports the actionability spectrum, realized through algorithmic issue detection and visual annotation generation. In a between-subjects study (N=36), novices revised a design under one of three conditions: textbook-based, awareness-centered, or solution-centered feedback. We found that solution-centered feedback led to fewer design issues and higher self-perceived creativity compared with textbook-based feedback, although expert ratings on creativity showed no significant differences. We discuss the implications for AI in Creativity Support Tools, including the potential of calibrating feedback actionability to help novices balance productivity with learning, growth, and developing design awareness.2026MLMingyi Li et al.Northeastern UniversityGenerative AI (Text, Image, Music, Video)Creative Collaboration & Feedback SystemsCHI
Characterizing User-Reported Risks across LLM ChatbotsAs Large Language Models (LLMs) become increasingly integral to daily life, users are engaging with multiple LLM chatbots for various needs; however, prior research on LLM risks often remains lab-based or focuses on single LLMs like ChatGPT or singular risks like privacy. To gain a multi-risk, cross-chatbot understanding of user experiences, we analyze Reddit discussions around seven major LLM chatbots using the NIST AI Risk Management Framework. We find that user-reported risks are unevenly distributed and chatbot-specific: ChatGPT is associated with safety and fairness concerns, Gemini with privacy, and Claude with security and resilience. Less frequent risks, such as explainability and privacy, appear as user trade-offs, whereas prevalent risks like fairness are experienced as direct harms. Our findings underscore the need to operationalize chatbot-specific risk mitigation, moving beyond system-centered risk mitigation to human-centered interventions that align with users' lived experiences.2026LLLingyao Li et al.University of South FloridaHuman-LLM CollaborationExplainable AI (XAI)AI Ethics, Fairness & AccountabilityCHI
Preshaping Hand Behaviour for Direct and Indirect Manipulation of 3D ObjectsEffortless manipulation informs and relies on preshaping: the subconscious posing of the hand before grasping. Virtual environments and the design of interaction techniques alters interaction requirements like contact and reach, forcing behavioural adaptation. We present a comparative study investigating preshaping behaviour across direct versus indirect (gaze-assisted) and bare-hand versus controller techniques on a docking task. Results reveal that response patterns scale with anticipated task difficulty, and that direct techniques elicit effective posing of the hand. Indirect techniques shortcut hand transport and, in turn lacks the sensory feedback to guide planning, inducing efficient but attenuated responses that necessitate compensatory manipulation and clutching. Notably, controllers that afford in-hand rotation allow users to extend their range of motion. These findings can inform interaction design to better afford preshaping and optimise 3D manipulation tasks.2026TMThorbjørn Mikkelsen et al.Aarhus UniversityHand Gesture RecognitionFull-Body Interaction & Embodied Input3D Modeling & AnimationCHI
Navigating Automated Hiring: Perceptions, Strategy Use, and Outcomes Among Young Job SeekersAs the use of automated employment decision tools (AEDTs) has rapidly increased in hiring contexts, especially for computing jobs, there is still limited work on applicants' perceptions of these emerging tools, and their experiences navigating them. To investigate, we conducted a survey with 448 computer science students -- young, current technology job-seekers -- about perceptions of the procedural fairness of AEDTs, their willingness to be evaluated by different AEDTs, the strategies they use relating to automation in the hiring process, and their job seeking success. We find that young job seekers' procedural fairness perceptions of and willingness to be evaluated by AEDTs varied with the level of automation involved in the AEDT, the technical nature of the task being evaluated, and their own use of strategies, such as job referrals. Examining the relationship of their strategies with job outcomes, notably, we find that referrals and family household income have significant and positive impacts on hiring success, while more egalitarian strategies (using free online coding assessment practice or adding keywords to resumes) did not. Overall, our work speaks to young job seekers' distrust of automation in hiring contexts, as well as the continued role of social and socioeconomic privilege in job seeking, despite the use of AEDTs that promise to make hiring "unbiased.''2025LALena Armstrong et al.Facilitating Equity and Fairness in TechCSCW
Communication Patterns Predict Team Skill in Multiplayer Online GamesThe present research on team collaboration is typically performed through qualitative interview based studies or social network measurements of connectedness through co-play. In this study, we take the unique approach to build networks from direct messages between players in the massive online game World of Tanks within and between games. Players self-organize into clans with specific roles assigned from military rankings (from private to commander). We explore the relationship between team communication volume and skill level, the impact of communication features on clan rating, and the differences in communication hierarchy within high versus low-rated clans. Our findings reveal that higher-rated clans engage in more pre-battle planning, suggesting that effective communication and strategic planning are key to team performance. Evidence shows teams who use voice chat during battle are significantly higher ranked. Finally, we reveal that the highest rated clans have more connected lower-ranked members emphasizing that these teams are "only as strong as their weakest link." This research is guided by the Transactive Memory Systems and Collective Intelligence theories which serve to expand the contribution of this research outside of games to other forms of virtual collaboration.2025ABAlexander J Bisberg et al.Online Interactions with Friends and StrangersCSCW
FEDT: Supporting Experiment Design and Execution in HCI Fabrication ResearchFabrication research in HCI relies on diverse experiments to inform and assess research contributions. However, performance and reporting of these experiments is inconsistent, not only reducing transparency that reassures reviewers and readers of a project's rigour but also challenging methods' replicability by future researchers. We analyze recent fabrication publications to extract a unified experimental workflow, which we develop into a domain-specific language, and into the openly-available Fabrication Experiment Design Tool (FEDT). FEDT facilitates designing and executing HCI fabrication experiments. We demonstrate its comprehensiveness by using FEDT to model 42 fabrication experiments from 10 papers, which leverage varied fabrication technologies and techniques, including requiring human intervention in their steps. We discuss FEDT and our modeled experiments with the papers' original authors to evaluate its precision and utility in real workflows, and we demonstrate functionality with end-to-end replications of two published experiments.2025VSValkyrie Savage et al.Prototyping & User TestingComputational Methods in HCIResearch Ethics & Open ScienceUIST
Answering Developer Questions with Annotated Agent-Discovered Program TracesDevelopers often find themselves asking questions that cut across a code base. Answering these questions requires gathering relevant facts and tracing flow through the program. Yet today’s tools offer limited support for answering these questions. Developers can either use imprecise AI tools that ignore flow or flow-tracing tools that impose a great number of choices. In this paper, we introduce a new kind of tool that answers questions better by bringing together elements of both AI and flow. We instantiate this idea in Trailblazer, a system underpinned by an AI agent that simulates an information forager, iteratively tracing program dependencies in search of answers. Then, Trailblazer packages information it found into an answer digest, which includes interactive, annotated traces of exploration. These traces can be stepped through to help developers orient to the code and find where the answer is distributed within it. In a lab study, Trailblazer helped participants answer questions more efficiently and gain greater familiarity with program flow than an AI question answering baseline. This shows how AI agents can leverage program flow to bring additional structure and clarity to its answers.2025LYLitao Yan et al.Identity & Avatars in XRHuman-LLM CollaborationComputational Methods in HCIUIST
FreeForm: Flexibly Augmenting Formulas with Synchronized Markup and Graphical EditsAuthors of typeset formulas augment those formulas to make them easier to understand. When they do so, they trade off between using markup tools like LaTeX and formula-unaware graphical editors. In this paper, we explore how editing tools could combine the best affordances of both kinds of tools. We develop FreeForm, a projectional editor wherein authors can augment formulas---with color, labels, spacing, and more---across multiple synchronized representations. Augmentations are created graphically using direct selections and compact menus. Those augmentations propagate to LaTeX markup, which can itself be edited and easily exported. In two lab studies, we observe the value of our editor versus baselines of a widely-used LaTeX document editor and a state-of-the-art formula augmentation tool. Finally, we make recommendations for the design of projectional markup augmentation editors.2025JTJeffrey Tao et al.University of Pennsylvania, Department of Computer and Information ScienceGraphic Design & Typography ToolsCustomizable & Personalized ObjectsCHI
The Unintended Costs of Platform Interventions: Black-Owned Restaurants and Yelp ReputationIn Spring 2020, digital review-based platform Yelp added the searchable ``Black-owned'' attribute to support Black-owned businesses. Based on the literature, the impacts of this design intervention were mixed. As such, we sourced an original dataset of 250,000+ Yelp reviews from Black and non-Black-owned restaurants in Detroit and Los Angeles. Performing statistical and trend analyses, we compared the reputation metrics of Black-owned restaurants to their non-Black-owned counterparts before and after the intervention. Although Yelp reported positive impacts, our results contribute to the growing evidence of the harms and unintended costs of platform interventions. Specifically, while awareness of Black ownership and the number of Black-owned restaurant reviews increased, assumedly among and by Yelp’s predominately non-Black users, Black-owned restaurants saw a decline in average star ratings. Altogether, the findings highlight the need to interrogate underlying assumptions in the design process, integrating critical race concepts to better contextualize and evaluate interventions targeting marginalized users.2025CMCameron Moy et al.University of Pennsylvania, Annenberg School for CommunicationContent Moderation & Platform GovernanceGender & Race Issues in HCICHI
Classroom Simulacra: Building Contextual Student Generative Agents in Online Education for Learning Behavioral SimulationStudent simulation supports educators to improve teaching by interacting with virtual students. However, most existing approaches ignore the modulation effects of course materials because of two challenges: the lack of datasets with granularly annotated course materials, and the limitation of existing simulation models in processing extremely long textual data. To solve the challenges, we first run a 6-week education workshop from N = 60 students to collect fine-grained data using a custom built online education system, which logs students' learning behaviors as they interact with lecture materials over time. Second, we propose a transferable iterative reflection (TIR) module that augments both prompting-based and finetuning-based large language models (LLMs) for simulating learning behaviors. Our comprehensive experiments show that TIR enables the LLMs to perform more accurate student simulation than classical deep learning models, even with limited demonstration data. Our TIR approach better captures the granular dynamism of learning performance and inter-student correlations in classrooms, paving the way towards a ``digital twin'' for online education.2025SXSonglin Xu et al.University of California San Diego, Department of Electrical and Computer EngineeringHuman-LLM CollaborationIntelligent Tutoring Systems & Learning AnalyticsCHI
PeerEdu: Bootstrapping Online Learning Behaviors via Asynchronous Area of Interest Sharing from Peer GazeHuman visual attention is susceptible to social influences. In education, peer effects impact student learning, but their precise role in modulating attention remains unclear. To this end, we have developed an online education system that provides visual feedback to students based on the area of interest sharing of peer students' gaze patterns. Our experiment (N=311) suggested that although peer attention manipulated students' gaze, individuals adapted their viewing strategies rather than always mirroring peer focus. Furthermore, intentionally guiding students' gaze along the lecture pace did not always improve learning outcomes. Instead, students able to adaptively adjust their focus based on personal needs showed enhanced performance. These findings elucidate how peer visual attention shapes students' gaze patterns, deepening understanding of peer influence on learning. They also offer insights into designing adaptive online learning interventions leveraging peer attention modelling to optimize student attentiveness and learning success.2025SXSonglin Xu et al.University of California San Diego, Department of Electrical and Computer EngineeringOnline Learning & MOOC PlatformsCollaborative Learning & Peer TeachingCHI
Interactions Beyond the Pandemic: Lessons Learned from Large-scale Emergency Remote Teaching in Higher EducationOnline education — given the enhanced access for diverse populations and flexible participation — has been a topic of interest for many computer science and learning science researchers. The sudden shift to online settings during the COVID-19 Emergency Remote Teaching (ERT) provided a valuable opportunity to examine the use of educational technologies on a global scale with various digital readiness skills, beyond many past works that relied on small lab studies. Following a PRISMA-inspired methodology grounded on Moore’s three types of classroom interaction, this descriptive review investigates 22 empirical research papers published during the COVID-19 ERT era focused on higher-education online classrooms. We explore the empirical evidence reported in the collected corpus, and given how ERT remains a likely future occurrence, we suggest key directions for future research, including a new learning paradigm that centralizes and augments Learner-Content interaction to balance between flexibility and structure of online learning.2025MYMatin Yarmand et al.University of California San Diego, Computer Science and Engineering; University of California San Diego, The Design LabOnline Learning & MOOC PlatformsCollaborative Learning & Peer TeachingCHI
Media Bias Detector: Designing and Implementing a Tool for Real-Time Selection and Framing Bias Analysis in News CoverageMainstream media, through their decisions on what to cover and how to frame the stories they cover, can mislead readers without using outright falsehoods. Therefore, it is crucial to have tools that expose these editorial choices underlying media bias. In this paper, we introduce the \mbd, a tool for researchers, journalists, and news consumers. By integrating large language models, we provide near real-time granular insights into the topics, tone, political lean, and facts of news articles aggregated to the publisher level. We assessed the tool’s impact by interviewing 13 experts from journalism, communications, and political science, revealing key insights into usability and functionality, practical applications, and AI's role in powering media bias tools. We explored this in more depth with a follow-up survey of 150 news consumers. This work highlights opportunities for AI-driven tools that empower users to critically engage with media content, particularly in politically charged environments.2025JWJenny S Wang et al.Harvard Business School, Harvard Business School; Microsoft ResearchGenerative AI (Text, Image, Music, Video)AI Ethics, Fairness & AccountabilityMisinformation & Fact-CheckingCHI
Generative AI and Perceptual Harms: Who’s Suspected of using LLMs?Large language models (LLMs) are increasingly integrated into a variety of writing tasks. While these tools can help people by generating ideas or producing higher quality work, like many other AI tools, they may risk causing a variety of harms, potentially disproportionately burdening historically marginalized groups. In this work, we introduce and evaluate perceptual harms, a term for the harms caused to users when others perceive or suspect them of using AI. We examined perceptual harms in three online experiments, each of which entailed participants evaluating write-ups from mock freelance writers. We asked participants to state whether they suspected the freelancers of using AI, to rank the quality of their writing, and to evaluate whether they should be hired. We found some support for perceptual harms against certain demographic groups. At the same time, perceptions of AI use negatively impacted writing evaluations and hiring outcomes across the board.2025KKKowe Kadoma et al.Cornell University, Information ScienceHuman-LLM CollaborationAI Ethics, Fairness & AccountabilityAlgorithmic Fairness & BiasCHI