Recent advances in generative AI have accelerated the deployment of virtual celebrity agents for commercial endorsements. However, little is known about how their visual style and vocal type, especially when these characteristics are generated via advanced AI reconstruction or voice cloning techniques, affect user trust, perceived realism, familiarity, and social presence. We extracted a representative clip of Sheldon Cooper from The Big Bang Theory as a baseline and generated multiple Sheldon virtual agents that varied in visual style (hand-sculpted, AI-reconstructed, or non-stylized) and vocal type (cloned or synthesized). A 3 × 2 within-subjects experiment (N = 30) revealed that visual style significantly affected user trust, perceived realism, familiarity, and social presence, while vocal type affected only perceived realism and familiarity. Comparisons between the virtual agents and the video baseline confirmed that even cutting-edge 3D modeling still differs significantly from authentic video representations. Behavioral data indicate that the relationship between interpersonal-distance variations and trust levels is not a simple linear one. These findings provide actionable guidance for designers leveraging generative AI to create trustworthy virtual avatars or agents and delineate new research avenues for virtual celebrity agents.

Interaction between humans and virtual agents is becoming increasingly prevalent. This research investigates the influence of virtual agents on users' opinions within a virtual stand-up comedy club setting, focusing on the socio-psychological mechanisms of emotional contagion and group conformity. We aimed to combine these two areas in a novel study approach. To this end, we designed and conducted an exploratory pilot study in which participants ($\mathrm{N}=20$) rated the funniness of jokes in VR, under varying conditions of virtual agent laughter (absence, medium intensity, strong intensity). Our analysis shows a significant positive linear effect on the log odds of higher funniness ratings when virtual agents exhibit laughter. Moreover, our study shows that this effect is mainly due to emotional contagion processes, individual differences in susceptibility to group conformity, and the interaction of both concepts, showing that emotional contagion happens only when the individual susceptibility to group conformity is high. The results offer valuable insights into the social impact of VR environments and the potential of virtual agents to shape user perception and opinion.

Conversational Virtual Agents (CVAs) offer a promising approach for enhancing user task performance in Mixed Reality (MR) environments. This paper explores the integration of a CVA into an MR application designed to assist in solving a 2D physical puzzle, offering enhanced spatial cognitive capabilities. Using the CVA classification architecture and the MiRAS (Mixed Reality Agents) Cube Taxonomy, we developed an MR system with a state-aware assistant to guide users in a puzzle-solving task only when requested. The primary research question is whether or not the CVA needs to be embodied. We conducted a study with 34 participants to investigate the influence of Voice-only and Embodied CVAs on puzzle-solving performance, user interactions with the assistant, the assistant’s social presence, overall system cognitive workload, and users’ perceptions of future system use. Both modalities showed equivalent outcomes regarding the number of position- and orientation-related queries, perceived usability, message and affective understanding, performance, frustration, and usefulness. However, results showed that Voice-only CVA significantly enhanced task efficiency: participants completed puzzles more quickly and accurately, reporting lower effort than in the Embodied condition. These findings suggest that Voice-only CVAs may be more effective for tasks like puzzle solving, where auditory guidance alone appears sufficient to support better performance.

With the continuous advancement of artificial intelligence technology, data-driven methods for reconstructing and animating virtual agents have achieved increasing levels of realism. However, there is limited research on how these novel data-driven methods, combined with voice cues, affect user perceptions. We use advanced data-driven methods to reconstruct stylized agents and combine them with synthesized voices to study their effects on users' trust and other perceptions (e.g. social presence and empathy). Through an experiment with 27 participants, our findings reveal that stylized virtual agents enhance user trust to a degree comparable to real style, while voice has a negligible effect on trust. Additionally, elder agents are more likely to be trusted. The style of the agents also plays a key role in participants' perceived realism, and audio-visual matching significantly enhances perceived empathy. These results provide new insights into designing trustworthy virtual agents and further support and validate the audio-visual integration theory.

Redirected walking is a locomotion technique that allows users to naturally walk through large immersive virtual environments (IVEs) by guiding them on paths that might vary from the paths they walk in the real world. While this technique enables the exploration of larger spaces in the IVE via natural walking, previous work has shown that it induces extra cognitive load for the users. On the other hand, previous research has shown that exposure to virtual nature environments can restore users’ diminished attentional capacities and lead to enhanced cognitive performances. Therefore, the aim of this paper is to investigate if the environment in which the user is redirected has the potential to reduce its cognitive demands. For this purpose, we conducted an experiment with 28 participants, who performed a spatial working memory task (i.e., 2-back test) while walking and being redirected with different gains in two different IVEs (i.e., (i) forest and (ii) urban). The results of frequentist and Bayesian analysis are consistent and provide evidence against an effect of the type of IVE on detection thresholds as well as cognitive and locomotion performances. Therefore, redirected walking is robust to the variation of the IVEs tested in this experiment. The results partly challenge previous research findings and, therefore, require future work in this direction.

Medical virtual agents (VAs) hold great potential to support patients in achieving their health goals, especially at times or in regions where the demand for physiological and psychological therapy exceeds the capacity of medical services. To create an accepted complement to on-site diagnosis, treatment, and counseling, it is critical to understand the impact of factors such as the agent’s visual representation, behavior, and responsibilities on creating a trustworthy human-agent relationship. To gain insights into these factors, we conducted a systematic literature review including 59 papers on embodied VAs in the medical domain. Our review focused on the application fields and the role of VAs in medicine, as well as the technology used to display them. Using thematic analysis, we discuss our findings in terms of user preferences, as well as potentials and barriers faced in the interaction with medical VAs. Concerning the visual representation, the users’ wish for customization in terms of appearance and communication modalities was pointed out. It was also important that the agent’s information builds up on trustworthy sources, that they are motivating and adapted to the users’ knowledge. Finally, our results identify research gaps, in particular regarding the technological implementation and the use of artificial intelligence.

Balance training has been shown to be effective in reducing risks of falling, which is a major concern for older adults. Usually, exercise programs are individually prescribed and monitored by physiotherapeutic or medical experts. Unfortunately, supervision and motivation of older adults during home-based exercises cannot be provided on a large scale, in particular, considering an ageing population. Augmented reality (AR) in combination with virtual coaches could provide a reasonable solution to this challenge.

Prior studies have explored the possibility of inducing social anxiety (SA) in virtual reality (VR). Among various existing protocols for this purpose, the Trier Social Stress Test (TSST) has been proven to be robust in evoking SA in the majority of participants in both in vivo as well as VR conditions. The TSST consists of giving a speech and performing mental arithmetic calculations each for five minutes in front of three persons. In this paper, we present an adaptation of TSST to investigate the effects of different numbers of virtual humans (VHs) (i.e., three, six, or fifteen) on perceived SA. In addition, we compare the results with an in vivo TSST with three real persons in the audience. Twenty four participants took part in this experiment. As a result, physiological arousal could be observed with VR inducing SA yet less than in vivo TSST. Furthermore, some of the subjective measures showed a high state of anxiety experienced during the experiment. An effect of the virtual audience size could be observed only in heart rate (HR) as a virtual audience size of three VHs induced the highest HR responses which was significantly different from an audience of size six and fifteen.

We report an experiment in which participants compared the size of a visual sphere to a haptic sphere in VR. The sizes from two modalities were either congruent or conflicting (with different disparities). Importantly, three standard references (small, medium and large) for haptic sizes were used with the method of constant stimuli. Results show a dominant functional priority of the visual size perception. Moreover, observers demonstrated a central tendency effect: over-estimation for smaller haptic sizes but under-estimation for larger haptic sizes. The results are discussed in the framework of adaptation level theory for haptic size reference, and provide important implications for the design of 3D visuo-haptic human-computer interaction.

We report on findings from two focus groups for designing an application for balance training at home with an augmented reality virtual coach. Following a User-Centered Design approach, we performed two focus groups with older adults at the early stages of development. Focus group participants were shown a prototype using a Meta 2 head mounted display. Their movements were tracked using a Kinect 2. The virtual coach gave balance training instructions and demonstrated their correct performance. Results suggest that, given the trade-offs of traditional health care, older adults are positive towards using an AR coach for their balance training.

The combination of virtual reality (VR) and gamification opens up new vistas for innovative forms of therapy for alcohol use disorder (AUD) and have enormous potential to improve traditional therapy methods. In this paper, three gamified and one non-gamified AUD therapy applications for VR are introduced and evaluated. The games are based on two behavioral therapy methods, which are Cue Exposure Therapy (CET) and Approach Avoidance Training (AAT). The games are realized in the context of a virtual supermarket, which is considered as a relapse-risky environment. The aim is to help AUD patients practice avoiding alcohol first in a VR-based simulation and later in a real supermarket. In preparation for a long-term clinical study, a usability study was conducted with 13 healthy participants. The results show that the VR game was enjoyed, increased the motivation, and fewer errors were made than in the comparable non-gamified application.

We evaluate the ability of locally present participants to localize an avatar head's gaze direction in 360° hosted telepresence. We performed a controlled user study to test two potential solutions to indicate a remote user's gaze. We analyze the influence of the user's distance to the avatar and display technique on localization accuracy. Our experimental results suggest that all these factors have a significant effect on the localization accuracy with varying effect sizes.