Advances in emerging technologies, such as on-body mechanical actuators and electrical muscle stimulation, have allowed computers to take control over our bodies. This presents opportunities as well as challenges, raising fundamental questions about agency and the role of our bodies when interacting with technology. To advance this research field as a whole, we brought together expert perspectives in a week-long seminar to articulate the grand challenges that should be tackled when it comes to the design of computers’ control over our bodies. These grand challenges span technical, design, user, and ethical aspects. By articulating these grand challenges, we aim to begin initiating a research agenda that positions bodily control not only as a technical feature but as a central, experiential, and ethical concern for future human–computer interaction endeavors.

Despite growing recognition that responsible AI requires domain knowledge, current work on conversational AI primarily draws on clinical expertise that prioritises diagnosis and intervention. However, much of everyday emotional support needs occur in non-clinical contexts, and therefore requires different conversational approaches. We examine how chaplains, who guide individuals through personal crises, grief, and reflection, perceive and engage with conversational AI. We recruited eighteen chaplains to build AI chatbots. While some chaplains viewed chatbots with cautious optimism, the majority expressed limitations of chatbots’ ability to support everyday well-being. Our analysis reveals how chaplains perceive their pastoral care duties and areas where AI chatbots fall short, along the themes of Listening, Connecting, Carrying, and Wanting. These themes resonate with the idea of attunement, recently highlighted as a relational lens for understanding the delicate experiences care technologies provide. This perspective informs chatbot design aimed at supporting well-being in non-clinical contexts.

Diverse teams tend to be more productive and creative and the same is said to hold for science, with evidence that diverse groups of authors publish in more impactful journals. Within HCI, there have been several investigations into the diversity of such author teams, but it is unclear whether more diverse groups of co-authors also produce more impactful HCI papers, and which diversity dimensions matter the most. We investigate this relationship using a dataset of all CHI papers and estimate gender, ethnicity, location, sector, and experience measures of diversity. Through a regression analysis on citations and awards, we determine that co-author diversity indeed affects paper impact, but only some forms of it. Where diversity in experience, ethnicity, sector, and location result in more impactful CHI papers, this does not look to be the case for gender diversity. Through follow-up analysis we further investigate three potential ways how author diversity could be influencing paper impact, by examining research area, novelty, and positionality of these papers. This shows that a paper’s research area also contributes to its impact, while varying by diversity dimensions, indicating a link between the two.

People are increasingly relying on technology for self-care, including, more recently, seeking help through conversational interfaces driven by large language models (LLMs). Yet, how interaction with LLMs has impacted people’s self-care processes is not well understood. Therefore, we collected 405 user stories posted on Reddit about using LLMs for self-care. We identified four key themes on how people use LLMs for this purpose: Letting go , Finding comfort , Building up , and Reflecting on . We interviewed twelve counsellors to capture their perspectives on this practice, given their professional expertise and understanding of healthy self-care practices. Our results show that counsellors recognised several benefits, such as using LLMs as stepping stones or springboards towards improved self-care. They also highlighted several areas of concern, such as unintended consequences that might negatively affect users. We discuss the dissonance around how the early adopters of LLMs appropriate this technology to care for themselves, how counsellors see such usage, and outline implications of using LLMs as a technology for self-care. • We present an analysis of 405 LLM self-care stories. • We used stories representative of themes in interviews with diverse counsellors. • We discuss user and counsellor perspectives on LLMs as self-care technologies.

Many people suffer from anxiety, commonly manifesting in avoidance behavior and stress. For example, they might not open letters because they suspect them to contain bad news. With the phone having taken on a central role in communication and service access, anxiety and avoidance also stem from and impact its use, such as when emails remain unread. However, while phone addition had been studied in depth, avoidance behavior related to phones has not. This necessitates a broader understanding in which ways anxiety impacts and results from phone use. To develop this understanding, we surveyed user stories on the anxiety they felt around interacting with their phones. From 197 stories, we identify eight different kinds of phone anxiety. We supplement this analysis in two ways: (1) a survey with 81 participants to analyze the external perceptions of phone anxiety, and (2) interviews of 12 individual users on their phone anxiety coping strategies.

Computing devices commonly act as tools, extending our abilities and shaping how we interact with the world. We investigate one such tool, the calculator, which helps with arithmetic, but also commonly offers specialized functions for conversions, formulas, or graphing. Through an analysis of calculator apps and use cases, we describe limitations of current calculators. Crucially, calculator apps remain detached from tasks, motivating us to explore how to more closely integrate calculation with the world through augmented reality (AR). AR calculators can directly use measurements and numbers from the world in calculations as well as display results of calculations in the world. We provide a conceptual account of calculation in AR, as well as video prototypes that concretize the concept across different scenarios. These examples demonstrate how moving tools like the calculator to AR offers tighter task integration and reduces the work required in translating between the world and computational tools.

Augmented Reality (AR) excels at altering what we see but non-visual sensations are difficult to augment. To augment non-visual sensations in AR, we draw on the visual language of comic books. Synthesizing comic studies, we create a design space describing how to use comic elements (e.g., onomatopoeia) to depict non-visual sensations (e.g., hearing). To demonstrate this design space, we built eight demos, such as speed lines to make a user think they are faster and smell lines to make a scent seem stronger. We evaluate these elements in a qualitative user study (N=20) where participants performed everyday tasks with comic elements added as augmentations. All participants stated feeling a change in perception for at least one sensation, with perceived changes detected by between four participants (touch) and 15 participants (hearing). The elements also had positive effects on emotion and user experience, even when participants did not feel changes in perception.

Users ask large language models (LLMs) to help with their homework, for lifestyle advice, or for support in making challenging decisions. Yet LLMs are often unable to fulfil these requests, either as a result of their technical inabilities or policies restricting their responses. To investigate the effect of LLMs denying user requests, we evaluate participants’ perceptions of different denial styles. We compare specific denial styles (baseline, factual, diverting, and opinionated) across two studies, respectively focusing on LLM’s technical limitations and their social policy restrictions. Our results indicate significant differences in users’ perceptions of the denials between the denial styles. The baseline denial, which provided participants with brief denials without any motivation, was rated significantly higher on frustration and significantly lower on usefulness, appropriateness, and relevance. In contrast, we found that participants generally appreciated the diverting denial style. We provide design recommendations for LLM denials that better meet peoples’ denial expectations.

Advances in generative text-to-image models are enabling new forms of personalized and adaptive media. We investigate the potential of such techniques through a generative adaptation of the fairy tale of Little Red Riding Hood. Specifically, we test two kinds of adaptations: (1) continuously adapting the visuals based on a face-to-emotion model, and (2) eliciting viewers’ fears and adapting the story accordingly. In either case, the adaptive versions are designed to make the story more scary and thus enhance the viewing experience in this dimension. We compare both variants against a baseline condition in a between-subjects study with 97 participants. Our results show that these adaptations significantly alter the viewing experience, modulated by viewers’ genre preferences.

Engaging with our devices as we engage with each other is problematic as it distracts us and diminishes our social interactions. Subtle interactions have been presented as an approach to reconcile personal and computing interactions, through less disrupting technology. Along those lines, we investigate showing information right on and next to the people we are engaging with. Body-based data visualization allows us to maintain our attention with others, but to also receive information at the same time. We explore potential designs of such body-based and especially on-face visualizations and create a set of five prototype visualizations in a Snapchat lens. We use these prototypes in a video call study with 16 participants to evaluate how body-based visualizations affect actual conversations.

Despite repeated reports of socially inappropriate and dangerous chatbot behaviour, chatbots are increasingly used as mental health services in providing support for young people. In sensitive settings as such, the notion of perceived moral agency (PMA) is crucial, given its critical role in human-human interactions. In this paper, we investigate the role of PMA in human-chatbot interactions. Specifically, we seek to understand how PMA influence the perception of trust, likeability, and perceived safety of chatbots for mental health across two distinct age groups. We conduct an online experiment(N = 279)to evaluate chatbots with low and high PMA as targeted towards teenagers and adults. Our results indicate increased trust, likeability, and perceived safety in mental health chatbots displaying high PMA. A qualitative analysis revealed four themes, assessing participants' expectations of mental health chatbots in general, as well as targeted towards teenagers: Anthropomorphism, Warmth, Sensitivity, and Appearance manifestation. We show that PMA plays a crucial role in influencing the perceptions of chatbots and provide recommendations for designing socially appropriate mental health chatbots.

OpenStreetMap (OSM) is a large online community where users collaborate to map the world. In addition to manual edits, the OSM mapping database is regularly modified by bots and automated edits. In this article, we seek to better understand how people and bots interact and conflict with each other. We start by analysing over 15 years of mailing list discussions related to bots and automated edits. From this data, we uncover five themes, including how automation results in power differentials between users and how community ideals of consensus clash with the realities of bot use. Subsequently, we surveyed OSM contributors on their experiences with bots and automated edits. We present findings about the current escalation and review mechanisms, as well as the lack of appropriate tools for evaluating and discussing bots. We discuss how OSM and similar communities could use these findings to better support collaboration between humans and bots.

Sense of control is increasingly used as a measure of quality in human-computer interaction. Control has been investigated mainly at a high level, using subjective questionnaire data, but also at a low level, using objective data on participants’ sense of agency. However, it remains unclear how differences in higher level, experienced control reflect lower level sense of control. We study that link in two experiments. In the first one we measure the low-level sense of agency with button, touchpad, and on-skin input. The results show a higher sense of agency with on-skin input. In the second experiment, participants played a simple game controlled with the same three inputs. We find that on-skin input results in both increased sense and experience of control compared to touchpad input. However, the corresponding difference is not found between on-skin and button input, whereas the button performed better in the experiment task. These results suggest that other factors of user experience spill over to the experienced control at rates that overcome differences in the sense of control. We discuss the implications for using subjective measures about the sense of control in evaluating qualities of interaction.

A compelling property of virtual reality is that it allows users to interact with objects as they would in the real world. However, such interactions are limited to space within reach. We present Poros, a system that allows users to rearrange space. After marking a portion of space, the distant marked space is mirrored in a nearby proxy. Thereby, users can arrange what is within their reachable space, making it easy to interact with multiple distant spaces as well as nearby objects. Proxies themselves become part of the scene and can be moved, rotated, scaled, or anchored to other objects. Furthermore, they can be used in a set of higher-level interactions such as alignment and action duplication. We show how Poros enables a variety of tasks and applications and also validate its effectiveness through an expert evaluation.

Spatial recordings allow viewers to move within them and freely choose their viewpoint. However, such recordings make it easy to miss events and difficult to follow moving objects when skipping through the recording. To alleviate these problems we present the Who Put That There system that allows users to navigate through time by directly manipulating objects in the scene. By selecting an object, the user can navigate to moments where the object changed. Users can also view trajectories of objects that changed location and directly manipulate them to navigate. We evaluated the system with a set of sensemaking questions in a think-aloud study. Participants understood the system and found it useful for finding events of interest, while being present and engaged in the recording.

Technological advances are enabling a new class of augmented reality (AR) applications that use bodies as substrates for input and output. In contrast to sensing and augmenting objects, body-based AR applications track people around the user and layer information on them. However, prototyping such applications is complex, time-consuming, and cumbersome, due to a lack of easily accessible tooling and infrastructure. We present Body LayARs, a toolkit for fast development of body-based AR prototypes. Instead of directly programming for a device, Body LayARs provides an extensible graphical programming environment with a device-independent runtime abstraction. We focus on face-based experiences for headset AR, and show how Body LayARs makes a range of body-based AR applications fast and easy to prototype.

We rely on our sight when manipulating objects. When objects are occluded, manipulation becomes difficult. Such occluded objects can be shown via augmented reality to re-enable visual guidance. However, it is unclear how to do so to best support object manipulation. We compare four views of occluded objects and their effect on performance and satisfaction across a set of everyday manipulation tasks of varying complexity. The best performing views were a see-through view and a displaced 3D view. The former enabled participants to observe the manipulated object through the occluder, while the latter showed the 3D view of the manipulated object offset from the object's real location. The worst performing view showed remote imagery from a simulated hand-mounted camera. Our results suggest that alignment of virtual objects with their real-world location is less important than an appropriate point-of-view and view stability.

Human-computer interaction is replete with ways of talking about qualities of interaction or interfaces, including if they are expressive, rich, fluid, or playful. An example of such a quality is subtle. While this word is frequently used in the literature, we lack a coherent account of what it means to be subtle, how to achieve subtleness in an interface, and what theoretical backing subtleness has. To create such an account, we analyze a sample of 55 publications that use the word subtle. We describe the variants of subtle interaction in the literature, including claimed benefits, empirical approaches, and ethical considerations. Not only does this create a basis for thinking about subtleness as a quality of interaction, it also works to show how to solidify varieties of quality in HCI. We conclude by outlining some open empirical and conceptual questions about subtleness.

Touch accuracy is not just dependent on the performance of the touch sensor itself. Instead, aspects like phone grip or occlusion of the screen have been shown to also have an influence on accuracy. Yet, these are all dependent on one underlying factor: the size and proportions of the user's hand. To better understand touch input, we investigate how 11 hand features influence accuracy. We find that thumb length in particular correlates significantly with touch accuracy and accounts for about 12% of touch error variance. Furthermore, we show that measures of some higher level interactions also correlate with hand size.

Grabbing users' attention is a fundamental aspect of interactive systems. However, there is a disconnect between the ways our devices notify us and how our bodies do so naturally. In this paper, we explore the body's modality of itching as a way to provide such natural feedback. We create itching sensations via low-current electric stimulation, which allows us to quickly generate this sensation on demand. In a first study we explore the design space around itching and how changes in stimulation parameters influence the resulting sensation. In a second study we compare vibration feedback and itching integrated in a smartwatch form factor. We find that we can consistently induce itching sensations and that these are perceived as more activating and interrupting than vibrotactile stimuli.

Current mobile devices commonly use vibration feedback to signal incoming notifications. However, vibration feedback exhibits strong attention capture, limiting its use to short periods and prominent notifications. Instead, we investigate the use of compression feedback for notifications, which scales from subtle stimuli to strong ones and can provide sustained stimuli over longer periods. Compression feedback utilizes inflatable straps around a user's limbs, a form factor allowing for easy integration into many common wearables. We explore technical aspects of compression feedback and investigate its psychophysical properties with several lab and in situ studies. Furthermore, we show how compression feedback enables reactive feedback. Here, deflation patterns are used to reveal further information on a user's query. We also compare compression and vibrotactile feedback and find that they have similar performance.

Emoji, a set of pictographic Unicode characters, have seen strong uptake over the last couple of years. All common mobile platforms and many desktop systems now support emoji entry, and users have embraced their use. Yet, we currently know very little about what makes for good emoji entry. While soft keyboards for text entry are well optimized, based on language and touch models, no such information exists to guide the design of emoji keyboards. In this article, we investigate of the problem of emoji entry, starting with a study of the current state of the emoji keyboard implementation in Android. To enable moving forward to novel emoji keyboard designs, we then explore a model for emoji similarity that is able to inform such designs. This semantic model is based on data from 21 million collected tweets containing emoji. We compare this model against a solely description-based model of emoji in a crowdsourced study. Our model shows good perfor mance in capturing detailed relationships between emoji.

With the increasing popularity of smartwatches over the last years, there has been a substantial interest in novel input methods for such small devices. However, feedback modalities for smartwatches have not seen the same level of interest. This is surprising, as one of the primary function of smartwatches is their use for notifications. It is the interrupting nature of current notifications on smartwatches that has also drawn some of the more critical responses to them. Here, we present a subtle notification mechanism for smartwatches that uses light scattering in a wearer's skin as a feedback modality. This does not disrupt the wearer in the same way as vibration feedback and also connects more naturally with the user's body.

Current soft keyboards for emoji entry all present emoji in the same way: in long lists, spread over several categories. While categories limit the number of emoji in each individual list, the overall number is still so large, that emoji entry is a challenging task. The task takes particularly long if users pick the wrong category when searching for an emoji. Instead, we propose a new zooming keyboard for emoji entry. Here, users can see all emoji at once, aiding in building spatial memory where related emoji are to be found. We compare our zooming emoji keyboard against the Google keyboard and find that our keyboard allows for 18% faster emoji entry, reducing the required time for one emoji from 15.6 s to 12.7 s. A preliminary longitudinal evaluation with three participants showed that emoji entry time over the duration of the study improved at up to 60 % to a final average of 7.5 s.

We present a novel way to recognize users by the way they press a button. Our approach allows low-effort and fast interaction without the need for augmenting the user or controlling the environment. It eschews privacy concerns of methods such as fingerprint scanning. Button pressing behavior is sufficiently discriminative to allow distinguishing users within small groups. This approach combines recognition and action in a single step, e.g., getting and tallying a coffee can be done with one button press. We deployed our system for 5 users over a period of 4 weeks and achieved recognition rates of 95% in the last week. We also ran a larger scale but short-term evaluation to investigate effects of group size and found that our method degrades gracefully for larger groups.

Many MOOCs report high drop off rates for their students. Among the factors reportedly contributing to this picture are lack of motivation, feelings of isolation, and lack of interactivity in MOOCs. This paper investigates the potential of gamification with social game elements for increasing retention and learning success. Students in our experiment showed a significant increase of 25% in retention period (videos watched) and 23% higher average scores when the course interface was gamified. Social game elements amplify this effect significantly -- students in this condition showed an increase of 50% in retention period and 40% higher average test scores.

Users often struggle to enter text accurately on touchscreen keyboards. To address this, we present a flexible decoder for touchscreen text entry that combines probabilistic touch models with a language model. We investigate two different touch models. The first touch model is based on a Gaussian Process regression approach and implicitly models the inherent uncertainty of the touching process. The second touch model allows users to explicitly control the uncertainty via touch pressure. Using the first model we show that the character error rate can be reduced by up to 7% over a baseline method, and by up to 1.3% over a leading commercial keyboard. Using the second model we demonstrate that providing users with control over input certainty reduces the amount of text users have to correct manually and increases the text entry rate.

For many people their phones have become their main everyday tool. While phones can fulfill many different roles they also require users to (1) make do with affordance not specialized for the specific task, and (2) closely engage with the device itself. We propose utilizing the space and objects around the phone to offer better task affordance and to create an opportunity for casual interactions. Such around-device devices are a class of interactors that do not require users to bring special tangibles, but repurpose items already found in the user's surroundings. In a survey study, we determine which places and objects are available to around-device devices. Furthermore, in an elicitation study, we observe what objects users would use for ten interactions.

We describe the focused-casual continuum, a framework for describing interaction techniques according to the degree to which they allow users to adapt how much attention and effort they choose to invest in an interaction conditioned on their current situation. Casual interactions are particularly appropriate in scenarios where full engagement with devices is frowned upon socially, is unsafe, physically challenging or too mentally taxing. Novel sensing approaches which go beyond direct touch enable wider use of casual interactions, which will often be 'around device' interactions. We consider the degree to which previous commercial products and research prototypes can be considered as fitting the focused-casual framework, and describe the properties using control theoretic concepts. In an experimental study we observe that users naturally apply more precise and more highly engaged interaction techniques when faced with a more challenging task and use more relaxed gestures in easier tasks.

We present imaginary reality games, i.e., games that mimic the respective real world sport, such as basketball or soccer, except that there is no visible ball. The ball is virtual and players learn about its position only from watching each other act and a small amount of occasional auditory feed-back, e.g., when a person is receiving the ball. Imaginary reality games maintain many of the properties of physical sports, such as unencumbered play, physical exertion, and immediate social interaction between players. At the same time, they allow introducing game elements from video games, such as power-ups, non-realistic physics, and player balancing. Most importantly, they create a new game dynamic around the notion of the invisible ball. To allow players to successfully interact with the invisible ball, we have created a physics engine that evaluates all plausible ball trajectories in parallel, allowing the game engine to select the trajectory that leads to the most enjoyable game play while still favoring skillful play.

Advances in sensing technology are currently bringing touch input to non-planar surfaces, ranging from spherical touch screens to prototypes the size and shape of a ping-pong ball. To help interface designers create usable interfaces on such devices, we determine how touch surface curvature affects targeting. We present a user study in which participants acquired targets on surfaces of different curvature and at locations of different slope. We find that surface convexity increases pointing accuracy, and in particular reduces the offset between the input point perceived by users and the input point sensed by the device. Concave surfaces, in contrast, are subject to larger error offsets. This is likely caused by how concave surfaces hug the user's finger, thus resulting in a larger contact area. The effect of slope on targeting, in contrast, is unexpected at first sight. Some targets located downhill from the user's perspective are subject to error offsets in the opposite direction from all others. This appears to be caused by participants acquiring these targets using a different finger posture that lets them monitor the position of their fingers more effectively.