On social networks, we experience false information spreading faster than truthful content, a dynamic increasingly amplified by the growing presence of AI-generated content. Users must assess authenticity on their own or with the support of emerging detection mechanisms. This paper presents the design and evaluation of a user-facing tool to detect AI-generated content on social networks, using X.com as an example. To examine how interaction design influences trust and perceived effectiveness, a comparative user study was conducted in which participants interacted with multiple design variants of the detection tool. Unlike prior work focused primarily on detection accuracy, this study emphasizes how interaction design influences trust, usability, and user acceptance of AI-assisted moderation systems. The results show that differences in presentation and interaction shape users’ perceptions of the tool, even when underlying capabilities remain unchanged. In particular, a blur-based warning approach was found to balance attention effectiveness with minimal disruption to the primary tasks of users. This design supports human–AI decision sharing by communicating uncertainty without overriding user agency. By foregrounding user experience and ethics-in-use, this work highlights limitations of accuracy-centered detection research and offers design insights for trustworthy combined intelligence systems in everyday social media contexts.

A motion-based virtual reality (VR) bicycle simulator allows for steering and pedaling as input, while visual information and platform movements are provided as output. The simulator offers a potential means of safely experiencing mountain biking (MTB); however, the impact of complex multidirectional tilting of the bicycle on user experience remains unclear. Therefore, the aim of this study was to investigate the effects of integrating tilt and pitch on users’ psychological experience during downhill cycling in a simulator. Twenty-one participants rode a simulator course designed to replicate a real MTB course and were instructed to pass through balls placed at ten turns (i.e., banks) along the course. Measurements were taken under two conditions: the nonmotion (NM) condition, in which the platform remained stationary, and the motion-based (M) condition, in which the platform moved in tilt and pitch according to the visual environment. After each condition, participants completed the Simulator Sickness Questionnaire (SSQ), the Igroup Presence Questionnaire (IPQ), and the short version of the User Experience Questionnaire (UEQ-S). Eighteen participants completed both conditions. There was no significant difference in the SSQ and the UEQ-S between the two conditions. In the IPQ, only the subscale spatial presence was significantly higher in the M condition than in the NM condition. The platform’s tilt and pitch movements during downhill cycling in a VR bicycle simulator had only a minimal impact on simulator sickness, presence, and user experience. These results indicate that a platform without motion may be sufficient for rehearsing MTB downhill courses for individuals with no prior MTB experience.

Virtual reality cycling simulators are increasingly used in research on urban mobility, rehabilitation, and road safety, yet their biomechanical fidelity remains under-validated due to the lack of standardized, publicly available datasets. We present a novel benchmark resource comprising synchronized full-body inertial measurement unit (IMU) data at 120 Hz and egocentric video from 10 participants who cycled an identical 1.4 km urban route in both Vienna and a motion-enabled VR simulator. Six body-worn sensors captured head, torso, arm, and leg movements, enabling detailed comparisons of pedaling rhythm, limb coordination, balance control, and visual attention patterns across environments. Our analyses reveal that VR successfully replicates fundamental locomotor patterns including pedaling cadence and bilateral leg coordination. However, significant differences emerge in upper-body dynamics: participants exhibited greater torso rotational variability in VR compared to real-world cycling, suggesting altered balance strategies. Head movement patterns showed broader yaw angles during virtual riding, yet these differences did not correlate with simulator sickness or immersion ratings. Emotional assessments indicated lower enjoyment and higher frustration in VR, highlighting gaps in affective realism despite mechanical similarity. This openly available dataset provides a critical resource for validating simulator designs, developing personalized VR training systems, and advancing research in embodied locomotion. The dataset, analysis tools, and documentation are freely accessible at: https://street2simulator.de/.

For centuries, building tools has been how humans compensate for what biology denied them. The rise of artificial intelligence is the latest and most consequential chapter in that story, because for the first time automation reaches into cognitive domains we once thought were distinctively human. This paper asks what that shift means for the humans on the other side of the interface. We draw on Arnold Gehlen’s philosophical anthropology to frame tool-making as constitutive of human identity, and on Fitts’ HABA-MABA framework to trace how the division of cognitive labor between humans and machines has evolved. Five case studies from our own research illustrate how combined intelligence plays out in practice across different everyday contexts. Together, they point toward a single design imperative: keep humans genuinely in control while improving overall outcomes.

Wearable-based human activity recognition (HAR) has emerged as a valuable method for capturing activities across diverse domains, including rehabilitation, occupational ergonomics, sports, and human-computer interaction (HCI). While recognition performance can be significantly enhanced by leveraging multiple complementary sensors, this approach requires accurately synchronized time bases across all devices. Although previous studies on synchronization in HAR suggested that sub-second accuracy is advisable while sub-100 ms accuracy is unnecessary, the specific effect of time discrepancies on machine learning models has, so far, remained unexplored. We address this gap by introducing an experimental paradigm for systematically evaluating the impact of time discrepancies in multi-wearable HAR, which we evaluated in two experiments. In our first experiment, we use the example of multi-stage temporal convolutional networks (MS-TCN) for sequence-to-sequence action segmentation, simulating the time discrepancies of time offset and clock skew via rational resampling. Our evaluation spanned 30,025 training and validation runs across different model configurations, totaling over one million core-hours of computation. Our results reveal that time offsets larger than 167 ms should be avoided in training datasets, and offsets beyond 333 ms can already significantly degrade HAR performance for typical activities of daily living (ADLs). Subsequently, we performed a second experiment focusing on the impact of time offsets on inference in models trained on synchronized datasets. Our evaluation spanned temporal convolutional networks, LSTMs, and Transformer architectures across five architectural configurations, each with two different temporal input lengths. The results indicate that LSTMs for action segmentation are more robust to desynchronization, while other architectures exhibited a marked performance degradation beyond desynchronization offsets spanning 167 ms. Our findings have implications for the design and deployment of multi-wearable HAR systems and may extend to other multi-sensor contexts.

Excessive, absent-minded, and aimless use of a smartphone can have a negative impact on both physical and mental health, with the potential to develop into an addiction. Because compulsive behavior can be triggered by the mere presence of a smartphone, conventional screen time management apps often fail to address this issue. We conducted expert interviews with therapists that revealed the potential of a few hours of abstinence with smartphones. Consequently, our research aims to offer a tangible physical opt-out solution while maintaining users’ autonomy to effectively reduce screen time. We designed a “mindful box” with a locking mechanism to securely store away digital devices. The box includes features that encourage intentional use, including gentle wake-up, meditation support, and cheer lists. In a one-week study with seven participants, the mindful box significantly reduced average screen time (p = 0.016). Users praised its effectiveness in combating distractions and limiting excessive device use.

As robots enter collaborative workspaces, ensuring mutual understanding between human workers and robotic systems becomes a prerequisite for trust, safety, and efficiency. In this position paper, we draw on the cooperation scenario of the AIMotive project in which a human and a cobot jointly perform assembly tasks to argue for a structured approach to intent communication. Building on the Situation Awareness-based Agent Transparency (SAT) framework and the notion of task abstraction levels, we propose a multidimensional design space that maps intent content (SAT1, SAT3), planning horizon (operational to strategic), and modality (visual, auditory, haptic). We illustrate how this space can guide the design of multimodal communication strategies tailored to dynamic collaborative work contexts. With this paper, we lay the conceptual foundation for a future design toolkit aimed at supporting transparent human-robot interaction in the workplace. We highlight key open questions and design challenges, and propose a shared agenda for multimodal, adaptive, and trustworthy robotic collaboration in hybrid work environments.

As autonomous agents, from self-driving cars to virtual assistants, become increasingly present in everyday life, safe and effective collaboration depends on human understanding of agents' intentions. Current intent communication approaches are often rigid, agent-specific, and narrowly scoped, limiting their adaptability across tasks, environments, and user preferences. A key gap remains: existing models of what to communicate are rarely linked to systematic choices of how and when to communicate, preventing the development of generalizable, multi-modal strategies. In this paper, we introduce a multidimensional design space for intent communication structured along three dimensions: Transparency (what is communicated), Abstraction (when), and Modality (how). We apply this design space to three distinct human-agent collaboration scenarios: (a) bystander interaction, (b) cooperative tasks, and (c) shared control, demonstrating its capacity to generate adaptable, scalable, and cross-domain communication strategies. By bridging the gap between intent content and communication implementation, our design space provides a foundation for designing safer, more intuitive, and more transferable agent-human interactions.

Frequently, pervasive computing research involves numerous distributed and interconnected components that must work in concert. Setting up the required technical backbone is often demanding, time-consuming, tedious, and repetitive. Existing tools for developing, testing, and tuning data-processing pipelines to extract high-level control data from low-level sensor data rely on specialized interfaces that are nontransparent, complicated, and inflexible, requiring specific personnel, such as programmers, which limits accessibility for specialists in other related fields, such as interaction designers. To address these challenges, we present a virtualization platform for data processing, transfer, and relaying. By coupling the visual programming paradigm with coherent data visualization and a network-based architecture, it facilitates adaptability and modularity, and is also compatible with a variety of hardware devices and third-party software.

Wearable-based human activity recognition (HAR) has become a relevant tool for identifying everyday activities in various domains like healthcare, sports, and human-computer interaction (HCI). The classification performance can be improved by using multiple complementary sensors, which require accurately matched time bases. Although previous studies on synchronization in HAR suggested that sub-second accuracy is advisable while sub-100 ms accuracy is unnecessary, the specific effect of time discrepancies on machine learning models remained unexplored. We address this with an empirical evaluation of the impact of time discrepancies in multi-wearable HAR. We apply a systematic approach using the example of multi-stage temporal convolutional networks (MS-TCN) for action segmentation, simulating the time discrepancies of time offset and clock skew via rational resampling. Our evaluation spanned 30,025 training and validation runs across different model configurations, totaling over one million core-hours of computation. Our results reveal that time offsets larger than 150 ms should be avoided in training datasets, and offsets beyond 300 ms can already significantly degrade the HAR performance for typical activities of daily living (ADLs). The findings highlight the need for adequate synchronization of training datasets. Our findings have implications for the design and deployment of multi-wearable HAR systems and may extend to other multi-sensor contexts.

Social robots are employed as companions, helping in industrial and domestic environments. Adapting robots’ capabilities to user needs can be achieved through teaching from human demonstrations. However, the influence of robots’ preexisting proficiency and learning rate on human teachers’ self-efficacy and perception of the robots is underexplored. In this paper, we simulated four robot performance types that combine: (1) preexisting proficiency (low/high) and (2) learning rate (slow/fast). We conducted a controlled lab experiment studying the impact of robots’ performance type on teachers’ self-efficacy, willingness to teach the robot, and perception of the robot (N=24), in which robots placed objects in suitable locations. Fast learners were perceived as more intelligent, anthropomorphic, and likable, and this caused higher teaching self-efficacy regardless of preexisting skills. Slow learners caused frustration while teaching. Moreover, participants stopped teaching robots with low preexisting skills sooner, regardless of the learning rate, indicating potential bias caused by expectations.

We present a fully textile capacitive touch sensor that provides an additional electrode for implementing driven (or active) shielding, which can considerably improve signal-to-noise ratio (SNR) and guard from parasitic capacitance. While driven shields are state-of-the-art for traditional printed sensors, they are still uncommon in contemporary textile user interfaces. Using an enameled copper wire as a bobbin thread in computerized machine embroidery, both sensor and shield electrodes are applied in a single sequence, eliminating manual intermediate or finishing steps and harnessing the design flexibility provided by the embroidery technique. Preliminary finite element analysis indicates significant improvements gained by employing driven shields, in terms of sensing range as well as signal quality, when the sensor is close to other conductors, e.g., when worn on the user’s body. By varying the shield electrode’s pattern properties, such as pattern type, area, and density, we investigate their effects on the resulting SNR, based on characterizations within controlled lab experiments. A major finding of our work is that the impact of density seems minor, while adjustments of the pattern layout seem to adequately compensate for a lower stitch density, with a grid layout yielding the best results.

In the digital age, technology permeates every aspect of our lives, offering connectivity but also posing risks to our well-being due to overuse. The concept of "digital detox" has emerged as a response, with smartphone apps supporting this process, yet the potential of wearable tech like smartwatches is less explored. Our study develops and tests a smartwatch-integrated digital detox aid, aiming to seamlessly blend with tech ecosystems offering a holistic solution. A preliminary mixed method user study (n=6) over two weeks assessed its efficacy in cutting down phone usage and app screen time, alongside monitoring phone interactions and physiological data. Initial results showed a decrease in screen time, which diminished in the second week, suggesting participant resistance and the intervention’s perceived intrusiveness. Despite proving the concept’s feasibility, the need for more user-aligned intervention methods and technical enhancements is clear, pointing to areas for future improvement.

Climate change communication demands narratives that resonate widely. “Carbon Rebellion,” an Interactive Digital Narrative, exemplifies this by blending personal anecdotes with data-driven content, fostering deep engagement. This research highlights the power of narratives in articulating the gravity of climate change, emphasizing the need for collective action and policy-level interventions. By juxtaposing individual choices against the backdrop of broader societal implications, the narrative seeks to both inspire and empower individuals. The findings reveal that such narratives can elicit strong emotional responses, catalyze critical thinking, and drive proactive climate action. The narrative’s emphasis on community-based efforts highlights the pivotal role of collective endeavors in addressing this global challenge.

Recycling organic waste through cornpostirig or fermentation is a simple but, so far, underrated way to capture CO2. Its effective and qualitative production requires the strict separation of recyclable organic waste from residual waste, but the persistent lack of public awareness still results in a considerable amount of contamination. Impurities from just a single container inevitably reduce the quality and can result in the expensive and less sustainable energy recovery of entire truckloads. To avoid the incineration of valuable organic waste, the problem must be tackled at the producer's site by preventing the addition of unsuitable waste. We developed a Smart Trash Can that allows to unobtrusively take photos of real waste in the home. Over six months, a total of 450 photos were collected, which were then manually labeled and segmented according to the captured waste types. Based on the collected dataset, two machine -learning approaches for computer vision have been implemented. While the first one binarily classifies the images as either being pure organic or containing impurities with a mean accuracy of 90.35%, the second one detects and segments impurities in the images with a mean accuracy of 98.24 % and a mean intersection over union value of 96.43 %, With the presented system, not only can contaminated waste be separated from pure organic waste, but also could feedback, incentives, or nudges encourage the producer to separate their waste more carefully in the future.

Robots are expected to be integrated into human workspaces, which makes the development of effective and intuitive interaction crucial. While vision- and speech-based robot interfaces have been well studied, direct physical interaction has been less explored. However, HCI research has shown that direct manipulation interfaces provide more intuitive and satisfying user experiences, compared to other interaction modes. This work examines how built-in force/torque sensors in robots can facilitate direct manipulation through nudge-based interactions. We conducted a user study (N = 23) to compare this haptic approach with traditional touchscreen interfaces, focusing on workload, user experience, and usability. Our results show that haptic interactions are more engaging and intuitive but also more physically demanding compared to touchscreen interaction. These findings have implications for the design of physical human-robot interaction interfaces. Given the benefits of physical interaction highlighted in our study, we recommend that designers incorporate this interaction method for human-robot interaction, especially at close quarters.