注目の論文一覧

Responsiveness in large language model (LLM) applications is widely assumed to be critical, yet the impact of latency on user behavior and perception of output quality has not been systematically explored. We report a controlled experiment varying time-to-first-token latency (2, 9, 20 seconds) across two taxonomy-driven knowledge task types (Creation and Advice). Log analyses reveal that user interaction behaviors were robust to latency, yet varied by task type: Creation tasks elicited more frequent prompting than Advice tasks. In contrast, participants who experienced 2-second latencies rated the LLM’s outputs less thoughtful and useful than those who experienced 9- or 20-second latencies. Participants attributed delays to AI deliberation, though long waits occasionally shifted this interpretation toward frustration or concerns about reliability. Overall, this work demonstrates that latency is not simply a cost to reduce but a tunable design variable with ethical implications. We offer design strategies for enhancing human-LLM interaction.

As Artificial Intelligence (AI) conversational agents become widespread, people are increasingly using them for health information seeking. The use of off-the-shelf conversational agents for health information seeking could place high metacognitive demands (the need for extensive monitoring and control of one's own thought process) on individuals, which could compromise their experience of seeking health information. However, currently, the specific demands that arise while using conversational agents for health information seeking, and the strategies people use to cope with those demands, remain unknown. To address these gaps, we conducted a think-aloud study with 15 participants as they sought health information using our off-the-shelf AI conversational agent. We identified the metacognitive demands such systems impose, the strategies people adopt in response, and propose considerations for designing beyond off-the-shelf interfaces to reduce these demands and support better user experiences and affordances in health information seeking.

Systems' latency — the time between user input and system response — slows down the human-computer interaction loop. Several studies revealed negative objective and subjective effects of high latency, typically treating latency as a constant delay. Because latency varies significantly in practice, recent work also assessed the effects of large and sudden latency changes. In practice, however, latency variations are small but frequent. As the effects of such variations are unclear, we investigate how small latency variations (+/- 50 ms) affect users' performance and perceived task load for 2D target selection tasks with static and moving targets. For static targets, we found that latency variation causes significantly higher completion times and less efficient trajectories, however with small effect sizes. In contrast, we found no significant effects on any performance measure for moving targets. Our findings indicate that the effect of latency variation is generally very small and quickly disappears for non-trivial tasks.

In interpersonal interactions, individuals often exhibit avoidance behaviors toward others they find unpleasant, which can undermine the comfort of everyday social experiences. Existing human-computer interaction (HCI) research has primarily focused on promoting social connections, while support for avoidance-oriented social situations remains underexplored. To address this gap, we propose leveraging Diminished Reality (DR) technology to obscure perceptual cues of undesirable individuals. We designed and implemented a mixed reality prototype system and conducted experiments manipulating both the occlusion method and social distance. Results indicate that DR significantly reduces users' social anxiety and sense of social presence. Moreover, participants generally expressed positive attitudes toward usage intention and ethical considerations. This work extends HCI research on social comfort, shifting the focus from "facilitating connection" to "supporting avoidance".

Clinician skepticism toward opaque AI hinders adoption in high-stakes healthcare. We present AICare, an interactive and interpretable AI copilot for collaborative clinical decision-making. By analyzing longitudinal electronic health records, AICare grounds dynamic risk predictions in scrutable visualizations and LLM-driven diagnostic recommendations. Through a within-subjects counterbalanced study with 16 clinicians across nephrology and obstetrics, we comprehensively evaluated AICare using objective measures (task completion time and error rate), subjective assessments (NASA-TLX, SUS, and confidence ratings), and semi-structured interviews. Our findings indicate AICare's reduced cognitive workload. Beyond performance metrics, qualitative analysis reveals that trust is actively constructed through verification, with interaction strategies diverging by expertise: junior clinicians used the system as cognitive scaffolding to structure their analysis, while experts engaged in adversarial verification to challenge the AI's logic. This work offers design implications for creating AI systems that function as transparent partners, accommodating diverse reasoning styles to augment rather than replace clinical judgment.

Despite their ubiquity, wireless earbuds remain audio-centric due to size and power constraints. We present VueBuds, the first camera-integrated wireless earbuds for egocentric vision, capable of operating within stringent power and form-factor limits. Each VueBud embeds a camera into a Sony WF-1000XM3 to stream visual data over Bluetooth to a host device for on-device vision language model (VLM) processing. We show analytically and empirically that while each camera's field of view is partially occluded by the face, the combined binocular perspective provides comprehensive forward coverage. By integrating VueBuds with VLMs, we build an end-to-end system for real-time scene understanding, translation, visual reasoning, and text reading; all from low-resolution monochrome cameras drawing under 5mW through on-demand activation. Through online and in-person user studies with 90 participants, we compare VueBuds against smart glasses across 17 visual question-answering tasks, and show that our system achieves response quality on par with Ray-Ban Meta. Our work establishes low-power camera-equipped earbuds as a compelling platform for visual intelligence, bringing rapidly advancing VLM capabilities to one of the most ubiquitous wearable form factors.

The use of AI in product design during early creative phases raises questions about its long-term consequences. Concerns are that extended AI use might inhibit creative cognitive processes, especially in novice designers. The aim of this study is to contribute to ongoing research in creative cognition and creative support tools such as AI in design. We conducted an exploratory study with 61 undergraduate students to analyze design exploration in sketching versus AI concept generation. The results indicate that AI groups produced a higher quantity and variation of total ideas (including text-based ideas), while sketch groups generated more image-based ideas. It was inconclusive whether the final image concepts from both AI and sketch groups were more creative. Additionally, homogenization effects were observed in the AI groups. Moreover, while the evolution of the design intent was evident in students who sketched, the focus in AI groups appeared to shift towards the tool (AI), which we analyzed as different design space exploration (DSE) prompting styles.

Advancements in artificial intelligence are challenging current policy frameworks. Both the human-computer interaction (HCI) community and policymakers note that technologies are designed better when they take into account the impact on society, and that policies are more effective when they are grounded in technical knowledge. Design research can be a powerful lens to support policy design processes. Driven by the potential for design research in technology policy development, the Robot Policy Design Toolkit (RPDT) was designed to support forecasting of robot technology policy and facilitate policy design experiences through a speculative design approach, centering forecasting, compromise, and simplicity design principles. This paper introduces the toolkit's design, reveals insights from how technologists design policies around social robots, and provides reflections from technology policy experts on the value and potential for design research tools, such as the RPDT, in policymaking contexts.

Cognitively stimulating experiences are fundamental to supporting the welfare of zoo-housed animals. Puzzle-feeders are often initially engaging, but require frequent human intervention and often lack adaptability to support animals’ sustained cognitive engagement. We developed a modular adaptive puzzle-feeder designed to support user agency, independence, and multisensory feedback. The system was deployed over four weeks with an Arctic fox (\textit{Vulpes lagopus}) across progressive difficulty levels and piloted with two coatis (\textit{Nasua narica}). Combining HCI and animal science methodologies, we assessed (1) multisensory engagement, (2) changes in behavioral diversity and habitat utilization, (3) adaptation to puzzle complexity, and (4) impact on human stakeholders. Results show strong sustained engagement (46.5\% time-budget), increased behavioral diversity, habitat exploration, strategic problem-solving, and positive keeper and visitor reactions. This work highlights how technology can support animal welfare and visitor experience, and how mixed HCI and ethological methods enable holistic evaluation of enrichment and animal usership.

Privacy policies are long, complex, and rarely read, which limits their effectiveness in informed consent. We investigate scrollytelling, a scroll-driven narrative approach, as a privacy policy presentation format. We built a prototype that interleaves the full policy text with animated visuals to create a dynamic reading experience. In an online study (N=454), we compared our tool against text, two nutrition-label variants, and a standalone interactive visualization. Scrollytelling improved user experience over text, yielding higher engagement, lower cognitive load, greater willingness to adopt the format, and increased perceived clarity. It also matched other formats on comprehension accuracy and confidence, with only one nutrition-label variant performing slightly better. Changes in perceived understanding, transparency, and trust were small and statistically inconclusive. These findings suggest that scrollytelling can preserve comprehension while enhancing the experience of policy reading. We discuss design implications for accessible policy communication and identify directions for increasing transparency and user trust.

Managing one’s digital footprint is overwhelming, as it spans multiple platforms and involves countless context-dependent decisions. Recent advances in agentic AI offer ways forward by enabling holistic, contextual privacy-enhancing solutions. Building on this potential, we adopted a “human-as-the-unit” perspective and investigated users’ cross-context privacy challenges through 12 semi-structured interviews. Results reveal that people rely on ad hoc manual strategies while lacking comprehensive privacy controls, highlighting nine privacy-management challenges across applications, temporal contexts, and relationships. To explore solutions, we generated nine AI agent concepts and evaluated them via a speed-dating survey with 116 US participants. The three highest-ranked concepts were all post-sharing management tools with half or full agent autonomy, with users expressing greater trust in AI accuracy than in their own efforts. Our findings highlight a promising design space where users see AI agents bridging the fragments in privacy management, particularly through automated, comprehensive post-sharing remediation of users’ digital footprints.

With the prevalence of mobile data visualizations, there have been growing concerns about their privacy risks, especially shoulder surfing attacks. Inspired by prior research on visual illusion, we propose BAIT, a novel approach to automatically generate privacy-preserving visualizations by stacking a decoy visualization over a given visualization. It allows visualization owners at proximity to clearly discern the original visualization and makes shoulder surfers at a distance be misled by the decoy visualization, by adjusting different visual channels of a decoy visualization (e.g., shape, position, tilt, size, color and spatial frequency). We explicitly model human perception effect at different viewing distances to optimize the decoy visualization design. Privacy-preserving examples and two in-depth user studies demonstrate the effectiveness of BAIT in both controlled lab study and real-world scenarios.

Traditional estate planning practices enable people to provide their heirs access to the assets left behind but are often insufficient for the transfer and management of online accounts. To understand how estate planning practices could be improved, we conducted 21 semi-structured interviews with older adults in the United States that explored their practices, concerns, and needs regarding postmortem online account access and management. We encountered few formalized digital estate planning practices; many participants use their credential management practices—primarily pen-and-paper—to provide postmortem account access. How participants envision account transfer is motivated by trust in their current practices and in their heirs, while concerns regarding technology hinder adoption of new methods. Participants consistently prioritize accounts with financial assets, and expectations surrounding postmortem account management vary based on individual circumstances, with the common goal of reducing burdens on executors and heirs. Our results suggest the need for developing technical standardization and expert guidance for digital estate planning.

Cybersecurity anxiety captures the persistent worry, stress and perceived lack of control individuals experience when navigating digital threats. While prior research has examined privacy concerns, computer anxiety and related constructs, no validated instrument exists to specifically measure anxiety in cybersecurity contexts. We address this gap with the Cybersecurity Anxiety Scale (CybAS), a 15-item psychometric instrument developed through literature review, item generation, and multiple survey studies. CybAS consists of three factors: Present (current concerns), Future (anticipated threats), and Control (perceived control over outcomes). Analyses confirm strong reliability and validity, and the concise format makes CybAS suitable for both research and applied settings. Beyond measurement, CybAS offers HCI researchers a diagnostic framework for detecting misalignments between users’ mental models and security technologies, enabling the design of anxiety-aware security systems that directly address emotional barriers, bridging the gap between usability, trust, and security.

The sensation of a drink in the throat is a salient example of the internal bodily feelings that shape our eating experiences. Computationally modeling these sensations would enable their redesign and inform technologies that augment how we eat. However, methods for quantifying such subjective, internal states from objective cues remain underdeveloped. This paper introduces a computational approach to bridge this gap. A first study (N=31) models subjective ratings from laryngeal skin temperature and ingested volume, revealing distinct, individual Interoceptive Profiles. Informed by these findings, we developed a wearable device that provides thermal feedback to the larynx. A second study (N=20) demonstrates that this intervention can alter drink sensations, contingent on the user's sensory profile. Based on these findings, we highlight the potential of the larynx as a site for bidirectional interaction (sensing and modulating) and propose a novel approach for personalized sensory augmentation.

Large language models (LLMs) are increasingly integrated into daily life through conversational interfaces, processing user data via natural language inputs and exhibiting advanced reasoning capabilities, which raises new concerns about user control over privacy. While much research has focused on potential privacy risks, less attention has been paid to the data control mechanisms these platforms provide. This study examines six conversational LLM platforms, analyzing how they define and implement features for users to access, edit, delete, and share data. Our analysis reveals an emerging paradigm of data control in conversational LLM platforms, where user data is generated and derived through interaction itself, natural language enables flexible yet often ambiguous control, and multi-user interactions with shared data raise questions of co-ownership and governance. Based on these findings, we offer practical insights for platform developers, policymakers, and researchers to design more effective and usable privacy controls in LLM-powered conversational interactions.

Video game archaeology is a relatively new field. This can involve studying players through the traces they leave in digital game worlds, though only limited work of this kind exists. Furthermore, the potential of these methods to record ephemeral play experiences for preservation purposes has not been widely explored. We conducted an archaeological survey of five sites in Elden Ring, taking place directly before, during and after the release of a major expansion. We present what is, to our knowledge, the first collaborative autoethnography of an archaeological survey in a video game, reflecting on our recorded footage, notes and data. Through a diffractive analysis, we demonstrate the value of video game archaeology as a form of hauntological practice that allows for a deeper reflection on the knowledge production process, and in doing so contribute to the development of new interdisciplinary methodologies in HCI, archaeology and games research.

Large Language Models (LLMs) are typically evaluated through general or domain-specific benchmarks testing capabilities that often lack grounding in the lived realities of end users. Critical domains such as healthcare require evaluations that extend beyond artificial or simulated tasks to reflect the everyday needs, cultural practices, and nuanced contexts of communities. We propose Samiksha, a community-driven evaluation pipeline co-created with civil-society organizations (CSOs) and community members. Our approach enables scalable, automated benchmarking through a culturally aware, community-driven pipeline in which community feedback informs what to evaluate, how the benchmark is built, and how outputs are scored. We demonstrate this approach in the health domain in India. Our analysis highlights how current multilingual LLMs address nuanced community health queries, while also offering a scalable pathway for contextually grounded and inclusive LLM evaluation.

Parenting while serving as a frontline leader is uniquely stressful, yet little is known about how family responsibilities shape physiological stress in these roles. We followed emergency physicians and tactical police leaders, comparing parents of young children with non-parents across four days: one critical mission day, two standard workdays, and one non-workday. Using wearable sensing, expert activity labeling, and daily debriefs, we inferred stress only in sedentary epochs via a normalized-heart-rate method, with an HRV-based index as benchmark. Parents showed higher stress on workdays and non-workdays, but not on critical mission days, where attentional narrowing and strict device policies appear to suppress parenting-related differences. We contribute: (i) in-the-wild physiological evidence that parenthood amplifies stress mainly under permeable boundaries, (ii) a pragmatic stress-labeling pipeline for safety-critical settings, (iii) a configuration-based account linking boundaries, attention, and parenting, and (iv) design implications for stress-aware boundary management systems, supported by an open analysis repository.

This study investigates how avatar customization in virtual reality (VR) impacts trust formation between unacquainted individuals and how such trust transfers to subsequent face-to-face (FtF) meetings. A user study with 48 participants was conducted, where participants were assigned to either a ``Similar-Self'' condition, with avatars resembling their real-world appearance, or an ``Alt-Self'' condition, with creative avatars. The results showed that ``Similar-Self'' avatars led to higher initial integrity-based trust perceptions, though both avatar conditions exhibited similar trust growth during VR encounters. Trust carried over from VR to FtF with a brief recalibration period and ultimately increased beyond VR levels in FtF encounters. This research provides insights into how VR can support the development of trust in early-stage interactions and offers implications for Social VR platforms to better support trustworthy interactions across virtual-physical boundaries.

Modern user interfaces are complex composites, with elements originating from various sources, such as the operating system, apps, a web browser, or websites. We posit that security and privacy decisions can to some extent depend on users correctly identifying an element's source, a concept we term "surface attribution." Through two large-scale vignette-based surveys (N=4,400 and N=3,057), we present the first empirical measurement of this ability. We find that users struggle, correctly attributing UI source only 55% of the time on desktop and 53% on mobile. Familiarity and strong brand cues are associated with improved accuracy, whereas UI positioning, a long-held security design concept especially for browsers, has minimal impact. Furthermore, simply adding a "Security & Privacy" brand cue to Android permission prompts failed to improve attribution. These findings demonstrate a fundamental gap in users' mental models, indicating that relying on them to distinguish trusted UI is a fragile security paradigm.

Algorithmic dating applications mediate romance through an "algorithmic mirror," subjecting users to data-driven classifications that shape their self-perception. However, the specific strategies users employ to interpret and strategically manage this reflection remain underexplored. Understanding this dynamic is critical, as navigating the algorithmic gaze demands significant emotional labor and has profound implications for user agency and well-being. Through semi-structured interviews with 15 OkCupid users, I investigated this process of sense-making. I contribute a novel typology of three knowledge forms, Folk, Personal, and Academic, that users construct to redefine themselves against the algorithm. Theoretically, this paper frames the "algorithmic other" as a statistical counterpart to Mead's "generalized other," revealing a core "dual-audience dilemma" where users perform for both humans and machines. These findings inform the design of more transparent and contestable systems that better support user agency.

Progressive digitisation and adoption of artificial intelligence (AI) are reshaping welfare services in ways that risk compounding inequalities for disabled people. Globally, many governments present these reforms as beneficial--streamlining processes, reducing costs and eliminating delays. Yet digitisation and automation of welfare decision-making can deepen exclusion and erode human accountability. In response, this paper foregrounds the lived experiences of people with the communication disability aphasia in navigating digitised welfare and their perspectives on AI-automated futures. We report findings from a four-stage participatory design study involving eight workshops with 42 recruited co-designers. Reflexive thematic analysis identified five challenges: the cost of performing disability, geographies of inequity, navigating digital bureaucracy, the accessibility paradox and hostile design. Co-designers voiced concerns about AI-automation but envisioned inclusive future alternatives: AI dialogues that are patient, multimodal and supportive; welfare systems that are compassionate, transparent and retain human recourse; and infrastructures that are open, publicly governed and truthful.

Past research showed that software developers often require explicit instructions to implement security measures. With the rapid rise of AI assistant tools such as ChatGPT, it remains unclear whether AI assistance supports or undermines secure practices, whether explicit security instructions are still essential, and how developers behave without guidance. To investigate these research questions, we conducted a qualitative lab study with 21 computer science students and a quantitative online study with 80 freelance developers. We focused on secure password storage and asked participants to implement registration logic under four conditions: without instructions, with AI assistance, with security instructions, or with both AI assistance and security instructions. Our study reveals a clear behavioral shift: In our task, many participants relied on AI-assisted code generation for security-related tasks, often prioritizing convenience over security. However, explicit security-focused instructions can redirect this behavior toward secure outcomes, demonstrating that AI tools alone are insufficient without targeted guidance.

Technology offers new opportunities to support healthier food choices, particularly for individuals in low-income communities who face systemic barriers to obtaining nutritious, affordable groceries. We introduce a novel conceptual model of grocery planning that frames food purchasing as a multi-objective optimization problem that considers cost, nutrition components, and a consumer's personal dietary goals. Guided by Zimmerman’s model of Self-Regulated Learning and prior research on food agency, we designed the Food Information System, a planning tool that provides optimized product recommendations aligned with users’ goals by integrating store inventory, prices, and nutritional data. We evaluated our system in an eight-week within-subjects intervention with 55 participants from a food-insecure community, followed by focus group sessions. While overall Healthy Eating Index scores remained largely stable, participants reported improved nutritional awareness and greater perceived agency in planning and purchasing groceries. We discuss design implications to support food agency by promoting long-term food literacy and by enhancing autonomy in making food choices.

Toxic behavior is a problem in online gaming platforms such as League of Legends (LoL), undermining player well-being and fairness. Platforms increasingly optimize “engagement” without distinguishing between positive and negative participation. Drawing on dual-process theory, we ask when hostile interactions can become economically productive. In an explanatory sequential mixed-methods study with LoL players, Study 1 (N = 430) models how reflective, System~2 brand bonds (i.e., brand personality, brand involvement, brand engagement) and negatively valenced, System~1 reactive responses (self-reported toxic behavior) relate to in-game spending. Study 2 (N = 80) uses reflexive thematic analysis to show how players interpret, repair, and channel frustration and hostility through cosmetics, events, and progression systems. Across studies, toxic behavior is positively associated with self-reported purchases and partially transmits the association between reflective brand attachments and spending. We contribute a dual-pathways account of how governance and monetization infrastructures can fold harmful engagement into value extraction, and we outline critical design provocations for centrally governed, highly monetized platforms.

As Smart Home Personal Assistants (SPAs) evolve into social agents, understanding user privacy necessitates interpersonal communication frameworks, such as Privacy Boundary Theory (PBT). To ground our investigation, our three-phase preliminary study (1) identified transmission and sharing ranges as key boundary-related risk factors, (2) categorized relevant SPA functions and data types, and (3) analyzed commercial practices, revealing widespread data sharing and non-transparent safeguards. A subsequent mixed-methods study (N=412 survey, N=40 interviews among the survey participants) assessed users' perceived privacy risks across data types, transmission ranges and sharing ranges. Results demonstrate a significant, non-linear escalation in perceived risk when data crosses two critical boundaries: the `public network' (transmission) and `third parties' (sharing). This boundary effect holds across data types and demographics. Furthermore, risk perception is modulated by data attributes, and contextual privacy calculus. Conversely, anonymization show limited efficacy especially for third-party sharing, a finding attributed to user distrust. These findings empirically ground PBT in SPA context and inform design of boundary-aware privacy protection.

Dark patterns are design practices that undermine users' ability to make autonomous and informed choices in digital experiences. The EU Digital Services Act (DSA) seeks to protect users from such designs and their effects, with Article 25 prohibiting three autonomy violation types: deception, manipulation and distortion/impairment. Demonstrating such regulatory violations, however, requires design-oriented reasoning necessary to articulate why an observed design practice constitutes a specific autonomy violation type. This paper maps 59 known dark patterns onto the three autonomy violation types from the DSA and identifies eight new design factors which can help determine when a dark pattern violates autonomy. Our mapping of dark patterns to autonomy violations grounds ongoing regulatory debates in design while opening pathways for translational research that reimagines how HCI engages with the governance of design practices.

Users of Virtual Reality (VR) primarily sense their environment through audiovisual cues. The lack of haptic feedback on their body can make them unaware of virtual obstacles outside their field of view. This lack of sensing can cause the user to unknowingly penetrate virtual objects, breaking the scene’s plausibility and disrupting the experience of other users in the same virtual space. We propose a haptic belt that increases the user’s scene awareness by rendering signals of collisions and proximity to virtual objects around the user. In a user study, we show that the belt improves spatial awareness both in a fast, high-stress scenario where the user's attention is limited and during a relaxed experience where the belt is the only source of information. The belt enables users to move closer to obstacles while reducing unintended collisions

Understanding and engaging with privacy policies is crucial for online privacy, yet these documents remain notoriously complex and difficult to navigate. We present PRISMe, an interactive browser extension that combines LLM-based policy assessment with a dashboard and customizable chat interface, enabling users to skim quick overviews or explore policy details in depth while browsing. We conduct a user study (N=22) with participants of diverse privacy knowledge to investigate how users interpret the tool's explanations and how it shapes their engagement with privacy policies, identifying distinct interaction patterns. Participants valued the clear overviews and conversational depth, but flagged some issues, particularly adversarial robustness and hallucination risks. Thus, we investigate how a retrieval-augmented generation (RAG) approach can alleviate issues by re-running the chat queries from the study. Our findings surface design challenges as well as technical trade-offs, contributing actionable insights for developing future user-centered, trustworthy privacy policy analysis tools.

While memes enhance social interaction on social media, they can raise privacy and security concerns. Despite research on overtly toxic or unsafe memes, little attention has been given to users' experiences with seemingly safe memes and how contextual factors trigger privacy concerns. This study explores users’ comfort levels, influencing factors, underlying reasons for discomfort, and unmet needs when engaging with such memes. We first collected and analyzed 2,317 Reddit posts describing real-world meme experiences, then conducted an online survey with 324 participants to evaluate comfort across curated scenarios. Our findings reveal that perceived-safe memes can cause harm when shared inappropriately, with comfort shaped by content and context. Privacy concerns intensify with deeper involvement, strangers, and sensitive meme topics. We identified users' desire for consent and control in meme interactions. Based on our study, we make recommendations for users, developers of social media platforms and policymakers to address meme-related privacy and contextual concerns.

Artificial Intelligence for Social Good (AI4SG) is a growing area that explores AI's potential to address social issues, such as public health. Yet prior work has shown limited evidence of its tangible benefits for intended communities, and projects frequently face real-world deployment and sustainability challenges. While existing HCI literature on AI4SG initiatives primarily focuses on the mechanisms of funded projects and their outcomes, much less attention has been given to the upstream funding agendas that influence project approaches. In this work, we conducted a reflexive thematic analysis of 35 funding documents, representing about $410 million USD in total investments. We uncovered a spectrum of conceptual framings of AI4SG and the approaches that funding rhetoric promoted: from biasing towards technology capacities (more techno-centric) to emphasizing contextual understanding of the social problems at hand alongside technology capacities (more balanced). Drawing on our findings on how funding documents construct AI4SG, we offer recommendations for funders to embed more balanced approaches in future funding call designs. We further discuss implications for how the HCI community can positively shape AI4SG funding design processes.

Mobile permission decisions are often studied at the moment a permission request appears. However, our study shows that users’ choices are shaped much earlier, across a multi-stage journey that begins with app-need recognition and unfolds through app discovery, exploration, selection, installation, and first use. Drawing on interviews with 19 U.S.\ Android users, we map this process and identify four archetypal journeys that explain how early cues, such as discovery sources, app type, and social trust, shape later permission behavior. These insights align with theoretical models like Privacy Calculus, showing how users weigh perceived benefits and risks at each step, and complement Contextual Integrity theory, explaining how social norms and information flows shape expectations and constrain privacy agency across steps. We contribute an empirically grounded framework that clarifies why permission outcomes vary across contexts. Our results reframe mobile privacy as a sequential, path-dependent process, offering implications for future design and research.

Passwords suffer from major usability hurdles that foster insecure practices and undermine cybersecurity. Passkeys were introduced to address these issues, however, adoption remains low. Digital nudges offer a promising way to accelerate passkey adoption, yet research lacks empirical insight about when to nudge and which nudge types and designs are most effective. We therefore employed a mixed-methods approach to examine the impact of nudges on passkey adoption across five touchpoints in the digital user journey: During registration, login, account recovery, while in the settings menu, and during user activity. First, we conducted 15 expert interviews to identify candidate nudges and their design principles. We evaluate these nudges in a randomized controlled trial (RCT) with 3,680 participants on a commercial healthcare platform. Our results indicate that digital nudges can significantly increase passkey adoption when applied at the right touchpoints, encouraging users to move beyond passwords.

We present an experimental study that investigates how LLM-driven conversational AI tools might be weaponized to facilitate, exacerbate, or commoditize coercive control. Inspired by speculative design, we construct four scenarios that combine well-known coercive control tactics with the current capabilities of conversational AI tools. Then, we explore these scenarios via interactions with popular AI agents (ChatGPT, Gemini). We find that although AI tools refuse straightforward requests for harmful content, their guardrails can be circumvented via strategies such as gradual persuasion, splitting conversations, pre-prompting, and manipulating the AI agent's settings. Collectively, these strategies enable AI agents to be leveraged in ways that facilitate harassment, intimidation, gaslighting, monitoring, surveillance, and other coercive control tactics. To make these tools safer for everyone, we discuss opportunities for AI agents to resist being abused for coercive control via analysis of users’ conversational patterns, and ensuring that pre-programmed settings are clearly visible to prevent covert manipulation.

The role of trust within Human-Computer Interaction is being redefined. With the increasing omnipresence, autonomy, and opacity of technology, users often struggle to understand the capabilities and limitations of systems. In this article, we present the results of an empirical study designed to provide a practical, evidence-based interpretation of trust propensity assessment using the Human-Computer Trust Scale (HCTS). We outline the process used to develop a guideline for interpreting the instrument’s results and explain the rationale for our decisions, advocating for calibrating trust in technology within HCI. Our findings demonstrate that the HCTS is a promising tool for conducting an initial evaluation of propensity to trust, but that such an assessment requires reflection and interpretation that should be considered within the context of the interaction

Robots are increasingly expected to share elevators with people, yet little is known about the conditions shaping acceptance. We introduce the Robot Boarding Area (RBA)—a designated entry zone for robots—and examine how its availability and congestion affect user evaluations. In an online survey, acceptance sharply decreased once the RBA was occupied by any person or large object, even under moderate crowding. A VR experiment confirmed this pattern and further showed that participants preferred when robots refrained from boarding in crowded conditions compared to forcing entry. By formalizing the RBA as an acceptance criterion and demonstrating the value of adaptive skip strategies, this work identifies spatial availability and boarding behavior as central to socially acceptable robot deployment in elevators.

Generative artificial intelligence (AI) adds unprecedented scale to capabilities for self-presentation online that may diverge from one’s physical-world identity, thus potentially misinforming consent to intimate interactions, such as in online dating. Yet there is little empirical understanding of AI-generated self-presentation and (non-)disclosure to interaction partners. We present a qualitative survey of 113 online daters who used AI-generated content in their profiles or messages seen by in-person meeting partners. Findings show that generative AI is often used to fabricate attractive dating personalities through profile text and bios, with no relevance to one’s actual identity, and is seldom disclosed to meeting partners to avoid romantic rejection. Because sexual assault is defined by mis- or under-informed consent, the study positions generative AI as a potentially significant sexual assault risk factor through its use for presentation of non-physical traits that are influential to dating outcomes yet not readily identified as AI-generated upon meeting face-to-face. Content warning: this paper discusses forms of sexual violence including rape by deception.

Mobile phone numbers function as single keys to banking, government, and commerce, making the Subscriber Identity Module (SIM) a critical element of security. In April 2025, South Korea’s largest carrier experienced a SIM breach that compromised authentication keys and exposed nearly 27 million subscriber identifiers. We conducted semi-structured interviews with mental-model elicitation (N=33) to examine user awareness, responses, and understanding of SIM-based authentication. Results reveal a pronounced awareness–action gap: participants recognized the breach yet held incomplete mental models, perceived little personal risk, and rarely acted protectively, even when affected. Learned helplessness, reliance on carriers, and the invisibility of SIM shaped these passive responses. Brief educational interventions improved conceptual understanding but seldom produced lasting behavioral change. Our findings demonstrate how technical opacity and psychological factors jointly inhibit protective action and offer design implications for usable security, emphasizing interventions that realign users’ mental models with system risks to foster sustainable practices.

Hands are the chief appendage with which we manipulate the world around us, creating sounds as they go. As such, they are a rich source of information that computers can leverage for input and context sensing. Indeed, many prior works in HCI have explored this idea by instrumenting users' hands with a microphone, often integrated into a ring, wristband, or watch. In this work, we explore an alternative bare-hands approach --- by using a microphone array integrated into a user's headset/glasses, we can use beamforming to create a virtual microphone that tracks with the user's fingers in 3D space. We show this method can capture even the subtle noise of a finger translating across surfaces, including skin-to-skin contact for micro-gestures, as well as passive widget interactions.

The dark patterns, deceptive interface designs manipulating user behaviors, have been extensively studied for their effects on human decision-making and autonomy. Yet, with the rising prominence of LLM-powered GUI agents that automate tasks from high-level intents, understanding how dark patterns affect agents is increasingly important. We present a two-phase empirical study examining how agents, human participants, and human-AI teams respond to 16 types of dark patterns across diverse scenarios. Phase 1 highlights that agents often fail to recognize dark patterns, and even when aware, prioritize task completion over protective action. Phase 2 revealed divergent failure modes: humans succumb due to cognitive shortcuts and habitual compliance, while agents falter from procedural blind spots. Human oversight improved avoidance but introduced costs such as attentional tunneling and cognitive load. Our findings show neither humans nor agents are uniformly resilient, and collaboration introduces new vulnerabilities, suggesting design needs for transparency, adjustable autonomy, and oversight.

While symbolic execution (SE) can discover software vulnerabilities, it has received limited practical adoption. A key barrier is that SE requires human expertise to understand the program’s state and prioritize paths to analyze. Traditionally, users controlled SE through programmatic API calls, but recent tooling now implements graphical user interfaces (GUI). However, it is unclear how these new features affect human-SE performance. To understand this impact, we conducted a controlled experiment where 24 vulnerability discovery experts were tasked with analyzing a binary using an SE tool with either API or GUI-based features. From this study, we identify (1) experts' SE process, and (2) the impact of GUI-based features on human-SE performance. Then we propose recommendations to improve SE tool design.

The rise of AI-powered tools like ChatGPT enables non-programmers to bypass programming screening questions, undermining internal validity in usable security and privacy, and software engineering studies. Past ChatGPT-resistant tasks proposed static visual questions, which ChatGPT can now circumvent. Therefore, we tested alternative approaches such as video- and audio-based screeners that reveal key information step by step under strict time constraints to distinguish programmers from non-programmers. To this end, we conducted a study with 74 participants across three groups: programmers, non-programmers without AI assistance, and non-programmers using ChatGPT. Our results showed that audio-based screeners were robust against ChatGPT-based cheating, as non-programmers struggled to find correct answers within time limits, whereas programmers demonstrated high accuracy with minimal time pressure. Based on our findings, we recommend six audio-based ChatGPT-resistant screening questions that maximize screening effectiveness and efficiency and suggest a 215-second instrument that includes 95.87% of programmers while excluding 99.69% of non-programmers.

Gathering information about AI systems is essential for contesting their use; it forms the basis of arguments about how AI is causing harm. Information thus plays a central role for advocates like lawyers, journalists, and auditors contesting harmful AI systems. However, there is little systematic understanding of how these actors, many of whom are newly encountering AI in their advocacy work, access and use information effectively in this process. Understanding this information work can offer valuable insights for supporting effective contestation of harmful AI systems. To better understand information work in AI contestation, we interviewed 18 advocates in the United States (US) who have contested the use of AI in high-stakes domains, such as public benefits and housing. We characterize advocates' strategies for accessing information that is useful for contestation, including a range of creative yet resource-intensive and risky workarounds that they use to overcome opacity. We discuss implications of our findings for the effectiveness of popular transparency policy strategies in the US and offer additional ways to support the social fabric that makes advocates' information work effective.

As automated vehicles (AVs) move toward mainstream adoption, understanding how users learn about and build trust in them is critical. Prior research shows that women hold safety concerns and report low trust and familiarity with AVs. While limited exposure is often cited as a cause, growing evidence indicates that women’s needs, preferences, and safety priorities remain insufficiently addressed in AV design and governance. We conducted ten dyadic and five individual semi-structured interviews with fifteen women, guided by feminist HCI principles. We then analysed findings through a socio-ecological framework to explore trust and learning. Our findings show that women's needs and expectations for AVs develop in conversation with gendered and caregiving responsibilities, and experiences of safety and vulnerability. Trust and learning co-evolve in this process as a dynamic association of forces influencing inclusive mobility. We contribute a feminist socio-ecological account of trust–learning dynamics, identifying design and policy interventions that support inclusive onboarding, institutional accountability, and community-based co-learning for equitable AV adoption.

Small and medium-sized enterprises (SMEs) are facing growing cybersecurity threats amidst limited resources and regulatory complexity. This complexity stems from diverse stakeholders in the regulatory process, including policymakers, industry associations, and companies that must implement the regulations. Misalignments between these different stakeholders can further compound the complexity. Against this backdrop, we investigate the cybersecurity mental models held by three stakeholder groups in Denmark’s defence sector and how these mental models might influence regulatory processes. Using a qualitative approach combining focus groups with 6 policymakers, 11 policy promoters (industry associations), and 12 policy implementers (SMEs), we reveal key misalignments in perceptions of risk, threats, cyber readiness, and policy interpretation. Our findings further show that SMEs often treat cybersecurity as a compliance task, while policymakers assume strategic readiness. Based on our results, we suggest recommendations for aligning governance frameworks with organisational realities.

Deceptive patterns, i.e. dark patterns and manipulative user interfaces (UI), are a widely used design method that aims to manipulate users to act against their own interests. These patterns may particularly influence people with less education, visual impairments, and older adults. Yet, access is a critical feature of the user experience (UX), development standards, and law. We considered whether and how the Web Content Accessibility Guidelines (WCAG) and related legislation, such as the European Accessibility Act (EAA), can act as a tool against deceptive patterns. We used these guidelines and legal statues in a heuristic evaluation to analyze whether and how deceptive patterns violate or conform to these standards. Although statistical analysis revealed no significant relationship, we identified three patterns implicated by the WCAG guidelines: Countdown Timer, Auto-Play, and Hidden Information. We offer this approach as one tool in the fight against UI-based deception and in support of inclusive design.

In response to calls for open data and growing privacy threats, organizations are increasingly adopting privacy-preserving techniques that add noise to published datasets. These techniques seek to protect privacy of data subjects while enabling useful analyses. With expert feedback, we developed empirically-driven documentation explaining the noise characteristics of two Wikipedia pageview datasets: one using rounding (heuristic privacy) and another using differential privacy (DP, formal privacy). We then used these documents to conduct a task-based contextual inquiry (n=15) exploring how data users—largely unfamiliar with these methods—perceive, interact with, and interpret privacy-preserving noise during data analysis. Participants readily used simple uncertainty metrics from the documentation, but struggled when computing confidence intervals across multiple noisy estimates. They better devised simulation-based approaches for computing uncertainty with DP-noised vs. rounded data. Surprisingly, several participants incorrectly believed DP's stronger utility implied weaker privacy protections. We offer design recommendations for documentation and tools to better support data users working with privacy-noised data.

The process of leaving high-pressure, identity-defining communities can produce profound identity changes. This leaving process propels some people to seek support online and to share their experiences publicly. We interviewed 13 social media content creators who made content as a part of, or in response to, their leaving process to understand their motivations and the ways audiences engaged with their work. We then explored how platforms transformed creators' work into collaborative spaces for social support. As creators gained audiences, their visibility introduced new incentives, obligations, and risks. Creators had to manage the challenges of maintaining safe spaces for their audiences, meeting audience expectations, and addressing heightened safety concerns for themselves. We end by discussing the networked structure of creator-centered communities, the impacts of platform on creator communities, and the emotional harms associated with being at the center of a community focused on social support.

As more people turn to chatbots for socioemotional support—often termed psychosocial AI—the stakes of understanding these interactions grow. Psychosocial AI might foster healthier human-human relationships—and also might exacerbate loneliness, abuse, and self-harm. We provide an empirical account of one less-studied facet: seeking AI advice on sex, dating, and relationships with other people. We recruited 25 people who use AI for relationship advice to a questionnaire, collecting 90 prompts illustrating their practices. Interviews with 17 further explored how they navigate AI’s limitations to achieve intimacy goals. Our findings detail (1) the roles that users imagine for AI in relationship advice; (2) how users navigate risks like sycophancy and overreliance to attain relational benefits; and (3) the folk theories users hold and the prompting tactics they employ to overcome AI’s limitations. We close with recommendations for human-AI interaction, AI safety, and sociotechnical research, towards AI that supports healthier digital intimacies.