As large language models (LLMs) become embedded in interactive text generation, disclosure of AI as a source depends on people remembering which ideas or texts came from themselves and which were created with AI. We investigate how accurately people remember the source of content when using AI. In a pre-registered experiment, 184 participants generated and elaborated on ideas both unaided and with an LLM-based chatbot. One week later, they were asked to identify the source (noAI vs withAI) of these ideas and texts. Our findings reveal a significant gap in memory: After AI use, the odds of correct attribution dropped, with the steepest decline in mixed human-AI workflows, where either the idea or elaboration was created with AI. We validated our results using a computational model of source memory. Discussing broader implications, we highlight the importance of considering source confusion in the design and use of interactive text generation technologies.

Generative AI tools for image creation are now mainstream, yet we know little about when and why observers judge them as “creative”. Previous human-robot interaction research suggests that revealing the creation process can raise perceived machine creativity and points that observer differences may moderate this effect. We take these observations from physical robots to the bigger domain of virtual text-to-image diffusion systems by manipulating perceptual evidence (PE), i.e., interface-visible cues about the generation process. We report two preregistered online experiments looking into PE and observer individual differences. Study 1 (N=298) used a within-subjects manipulation comparing Product (final image only) to Product+Process (adding a short animation of the denoising process). Study 2 (N=295) added a between-subjects tutorial (diffusion vs. control) in a 2 × 2 mixed design. The tutorial briefly explained how diffusion models generate images, intended to raise system-specific literacy. Contrary to previous work, confirmatory analyses found no average effect of showing Process on creativity, and no tutorial effect. Exploratory analyses revealed that general AI literacy moderated the PE contrast, i.e., at lower literacy, observing process tended to lower creativity ratings; at higher literacy, it tended to raise them. Moreover, attitudes toward AI and art interest were positively associated with creativity ratings. Thematic analysis of open-ended responses indicated potential reasons for the lack of overall PE effect. Taken together, these converging quantitative and qualitative findings indicate that individual differences systematically shape creativity judgments of text-to-image GenAI and, in our setting, exert stronger and more reliable influence than PE alone. For design, this implies that process visualizations could help some audiences more than others. Interfaces that adapt to literacy and attitudes, or that pair process views with contextual explanation calibrated to user background, could be more likely to shift judgments than one-size-fits-all depictions of generation.

People face challenges with complex search tasks daily; chat-based large language models have the potential to assist search by helping with query formulation and content navigation, for instance. However, there is a lack of user research on integrating AI chat and traditional search to enhance Information Retrieval (IR) for the general public. This study explores how younger (25 participants) and older (22 participants) adults utilize these two platforms to retrieve information. Through two IR tasks, quality- and quantity-oriented tasks, focused on gathering evidence related to various countries’ CO 2 emissions, younger users spent more time, issued more queries, and explored more external web pages than older adults in both tasks. In contrast, despite reporting less familiarity with AI chat, older adults relied more on chat to complete their tasks. They submitted longer queries for the quality task and reached the first query faster for the quantity task than their younger counterparts, reflecting a top-down IR strategy. In the end, task outcomes were similar for the two groups, indicating AI chat’s capability to support IR. Drawing from our study results and relevant literature, we propose five design requirements to harness the synergic value of search and AI chat in supporting IR.

Publisher Copyright: © 2025 Copyright held by the owner/author(s).

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.

Seamless transition between public dialogue and private talks is essential in everyday conversations. Social Virtual Reality (VR) has revolutionized interpersonal communication by creating a sense of closeness over distance through virtual avatars. However, existing social VR platforms are not successful in providing safety and supporting private conversations, thereby hindering self-disclosure and limiting the potential for meaningful experiences. We approach this problem by exploring the factors affecting private conversations in social VR applications, including the usability of different interaction methods and the awareness with respect to the virtual world. We conduct both expert interviews and a controlled experiment with a social VR prototype we realized. We then leverage the outcomes of the two studies to establish a design space that considers diverse dimensions (including privacy levels, social awareness, and modalities), laying the groundwork for more intuitive and meaningful experiences of private conversation in social VR.

Increasing encounters between people and autonomous service robots may lead to conflicts due to mismatches between human expectations and robot behaviour. This interactive online study (N = 335) investigated human-robot interactions at an elevator, focusing on the effect of communication and behavioural expectations on participants’ acceptance and compliance. Participants evaluated a humanoid delivery robot primed as either submissive or assertive. The robot either matched or violated these expectations by using a command or appeal to ask for priority and then entering either first or waiting for the next ride. The results highlight that robots are less accepted if they violate expectations by entering first or using a command. Interactions were more effective if participants expected an assertive robot which then asked politely for priority and entered first. The findings emphasize the importance of power expectations in human-robot conflicts for the robot’s evaluation and effectiveness in everyday situations.

Heightened AI expectations facilitate performance in human-AI interactions through placebo effects. While lowering expectations to control for placebo effects is advisable, overly negative expectations could induce nocebo effects. In a letter discrimination task, we informed participants that an AI would either increase or decrease their performance by adapting the interface, when in reality, no AI was present in any condition. A Bayesian analysis showed that participants had high expectations and performed descriptively better irrespective of the AI description when a sham-AI was present. Using cognitive modeling, we could trace this advantage back to participants gathering more information. A replication study verified that negative AI descriptions do not alter expectations, suggesting that performance expectations with AI are biased and robust to negative verbal descriptions. We discuss the impact of user expectations on AI interactions and evaluation.

For individuals with Social Anxiety (SA), interacting with others can be a challenging experience, a concern that extends into the virtual world. While technology has made significant strides in creating more realistic virtual human agents (VHA), the interplay of gaze and interpersonal distance when interacting with VHAs is often neglected. This paper investigates the effect of dynamic and static Gaze animations in VHAs on interpersonal distance and their relation to SA. A Bayesian analysis shows that static centered and dynamic centering gaze led participants to stand closer to VHAs than static averted and dynamic averting gaze, respectively. In the static gaze conditions, this pattern was found to be reversed in SA: participants with higher SA kept larger distances for static-centered gaze than for averted gaze VHAs. These findings update theory, elucidate how nuanced interactions with VHAs must be designed, and offer renewed guidelines for pleasant VHA interaction design.

While Mixed Reality allows the seamless blending of digital content in users' surroundings, it is unclear if its fusion with physical information impacts users' perceptual and cognitive resources differently. While the fusion of digital and physical objects provides numerous opportunities to present additional information, it also introduces undesirable side effects, such as split attention and increased visual complexity. We conducted a visual search study in three manifestations of mixed reality (Augmented Reality, Augmented Virtuality, Virtual Reality) to understand the effects of the environment on visual search behavior. We conducted a multimodal evaluation measuring Fixation-Related Potentials (FRPs), alongside eye tracking to assess search efficiency, attention allocation, and behavioral measures. Our findings indicate distinct patterns in FRPs and eye-tracking data that reflect varying cognitive demands across environments. Specifically, AR environments were associated with increased workload, as indicated by decreased FRP - P3 amplitudes and more scattered eye movement patterns, impairing users' ability to identify target information efficiently. Participants reported AR as the most demanding and distracting environment. These insights inform design implications for MR adaptive systems, emphasizing the need for interfaces that dynamically respond to user cognitive load based on physiological inputs.

Human-AI interaction in text production increases complexity in authorship. In two empirical studies (n1 = 30 & n2 = 96), we investigate authorship and ownership in human-AI collaboration for personalized language generation. We show an AI Ghostwriter Effect : Users do not consider themselves the owners and authors of AI-generated text but refrain from publicly declaring AI authorship. Personalization of AI-generated texts did not impact the AI Ghostwriter Effect , and higher levels of participants’ influence on texts increased their sense of ownership. Participants were more likely to attribute ownership to supposedly human ghostwriters than AI ghostwriters, resulting in a higher ownership-authorship discrepancy for human ghostwriters. Rationalizations for authorship in AI ghostwriters and human ghostwriters were similar. We discuss how our findings relate to psychological ownership and human-AI interaction to lay the foundations for adapting authorship frameworks and user interfaces in AI in text-generation tasks.

Biocybernetic loops encompass users' state detection and system adaptation based on physiological signals. Current adaptive systems limit the adaptation to task features such as task difficulty or multitasking demands. However, virtual reality allows the manipulation of task-irrelevant elements in the environment. We present a physiologically adaptive system that adjusts the virtual environment based on physiological arousal, i.e., electrodermal activity. We conducted a user study with our adaptive system in social virtual reality to verify improved performance. Here, participants completed an n-back task, and we adapted the visual complexity of the environment by changing the number of non-player characters. Our results show that an adaptive virtual reality can control users' comfort, performance, and workload by adapting the visual complexity based on physiological arousal. Thus, our physiologically adaptive system improves task performance and perceived workload. Finally, we embed our findings in physiological computing and discuss applications in various scenarios.

Virtual Reality (VR) games and experiences predominantly have the users interact while standing or seated. However, this only represents a fraction of the full diversity of human movement. In this paper, we explore a novel non-standing approach to VR locomotion where the user performs locomotion movements in the air or only slightly touching the ground with their feet. For instance, the user may lie supine on the ground, reminiscent of the Bicycle Crunch, a core training movement common in Pilates and other forms of bodyweight exercise. Although this cannot generally replace traditional VR locomotion, it provides two benefits that we believe can be of use for specific application domains such as VR exergames: First, the user's lower body movement is not impeded by a small real-life space, allowing versatile navigation of large virtual worlds using walking, running, strafing, and jumping. Second, we allow new ways to activate parts of the body that remain passive in most existing VR interactions. We describe and discuss four different variants of the approach, and investigate two prototypes further in a qualitative user study, to better understand their strengths, weaknesses, and application potential.

Prior research explored ways to alert virtual reality users of bystanders entering the play area from afar. However, in confined social settings like sharing a couch with seatmates, bystanders' proxemic cues, such as distance, are limited during interruptions, posing challenges for proxemic-aware systems. To address this, we investigated three visualizations, using a 2D animoji, a fully-rendered avatar, and their combination, to gradually share bystanders' orientation and location during interruptions. In a user study (N=22), participants played virtual reality games while responding to questions from their seatmates. We found that the avatar preserved game experiences yet did not support the fast identification of seatmates as the animoji did. Instead, users preferred the mixed visualization, where they found the seatmate's orientation cues instantly in their view and were gradually guided to the person's actual location. We discuss implications for fine-grained proxemic-aware virtual reality systems to support interaction in constrained social spaces.

Modern games make creative use of First- and Third-person perspectives (FPP and TPP) to allow the player to explore virtual worlds. Traditionally, FPP and TPP perspectives are seen as distinct concepts. Yet, Virtual Reality (VR) allows for flexibility in choosing perspectives. We introduce the notion of a perspective continuum in VR, which is technically related to the camera position and conceptually to how users perceive their environment in VR. A perspective continuum enables adapting and manipulating the sense of agency and involvement in the virtual world. This flexibility of perspectives broadens the design space of VR experiences through deliberately manipulating perception. In a study, we explore users’ attitudes, experiences and perceptions while controlling a virtual character from the two known perspectives. Statistical analysis of the empirical results shows the existence of a perspective continuum in VR. Our findings can be used to design experiences based on shifts of perception.

Social interactions are an essential part of many digital games, and provide benefits to players; however, problematic social interactions also lead to harm. To inform our understanding of the origins of harmful social behaviours in gaming contexts, we examine how trait psychopathy influences player perceptions and behaviours within a gaming task. After measuring participants’ (n=385) trait-level boldness, meanness, and disinhibition, we expose them to neutral and angry social interactions with a non-player character (NPC) in a gaming task and assess their perceptions, verbal responses, and movement behaviours. Our findings demonstrate that the traits significantly influence interpretation of NPC emotion, verbal responses to the NPC, and movement behaviours around the NPC. These insights can inform the design of social games and communities and can help designers and researchers better understand how social functioning translates into gaming contexts.

Augmented Reality (AR) embeds virtual content in physical spaces, including virtual agents that are known to exert a social presence on users. Existing design guidelines for AR rarely consider the social implications of an agent’s personal space (PS) and that it can impact user behavior and arousal. We report an experiment (N=54) where participants interacted with agents in an AR art gallery scenario. When participants approached six virtual agents (i.e., two males, two females, a humanoid robot, and a pillar) to ask for directions, we found that participants respected the agents’ PS and modulated interpersonal distances according to the human-like agents’ perceived gender. When participants were instructed to walk through the agents, we observed heightened skin-conductance levels that indicate physiological arousal. These results are discussed in terms of proxemic theory that result in design recommendations for implementing pervasive AR experiences with virtual agents.

Embedded interfaces support active indexing, the process of making sense of the experienced. While this supports the creation of meaningful relationships, findings are limited on how it could support sustaining the significance of historical places. We selected two historical places of daily use and developed two AR apps for each. The apps aimed at fostering participants’ indexing and meaningful connection to the places by presenting their history. One app offered passive, informative content. The other displayed an interactive, moral dilemma emphasizing the places’ social values. We tested the apps in two between-subjects studies with N=42 participants in total. We explored how participants ’ historical understanding can be improved by embedding our interfaces on five levels (spatially, physically, contextually, historically, and morally). We found that the morally embedded AR stories supported active indexing stronger than the passive ones and argued that they could foster meaningful relationships between users and historical places.

Private homes are increasingly becoming smart spaces. While smart homes promise comfort, they expose most intimate spaces to security and privacy risks. Unfortunately, most users today are not equipped with the right tools to assess the vulnerabilities or privacy practices of smart devices. Further, users might lose track of the devices installed in their homes or are unaware of devices placed by a partner or host. We developed SaferHome, an interactive digital-physical privacy framework, to provide smart home users with security and privacy assessments and a sense of device location. SaferHome includes a digital list view and physical and digital dashboards that map real floor plans. We evaluated SaferHome with eight households in the wild. We find that users adopted various strategies to integrate the dashboards into their understanding and interpretation of smart home privacy. We present implications for the design of future smart home privacy frameworks that are impacted by technical affinity, device types, device ownership, and tangibility of assessments.

In medicine, patients can obtain real benefits from a sham treatment. These benefits are known as the placebo effect. We report two experiments (Experiment I: N = 369; Experiment II: N = 100) demonstrating a placebo effect in adaptive interfaces. Participants were asked to solve word puzzles while being supported by no system or an adaptive AI interface. All participants experienced the same word puzzle difficulty and had no support from an AI throughout the experiments. Our results showed that the belief of receiving adaptive AI support increases expectations regarding the participant’s own task performance, sustained after interaction. These expectations were positively correlated to performance, as indicated by the number of solved word puzzles. We integrate our findings into technological acceptance theories and discuss implications for the future assessment of AI-based user interfaces and novel technologies. We argue that system descriptions can elicit placebo effects through user expectations biasing the results of user-centered studies.