The global prevalence of mental health disorders has created a substantial treatment gap. To support clinicians and increase access to care, researchers in the field of Artificial Intelligence (AI) and Virtual Reality (VR) have investigated technology-mediated psychotherapy for years. However, research about stakeholders' concerns and their readiness to use AI in psychotherapy remains scarce. This study focuses on a user-centered approach to accommodate patients' concerns and, based on the results, implement measures to foster self-disclosure and trust towards an embodied AI therapist in VR. First, we conducted an online study with mental health patients $(N=152)$, which identified data autonomy and transparency as their primary ethical concerns. In a subsequent in-person VR study $(N=90)$ we compared effects of increased data autonomy and transparency on self-disclosure and trust towards an embodied AI therapist. Results indicated that higher data autonomy led to greater self-disclosure, while transparency had no significant effect. Manipulating data autonomy and transparency did not affect perceived trust, though exploratory calculations revealed that women reported significantly higher trust levels than men. These findings illuminate patients' priorities and provide implications for technical designs for AI-driven mental health care.

Spatial computing is set to change the way we work. It will enable both focused work through a higher degree of immersion and collaborative work through enhanced integration of shared interaction spaces or interaction partners. With the Apple Vision Pro, the level of immersion can be adjusted seamlessly. So far, there have been no systematic studies on how this adjustability affects work performance when working alone or together. The present empirical study fills this research gap by varying the level of immersion across three stages (high, medium, low) while solving various tasks with the Apple Vision Pro. The results show that selective attention improves significantly with increasing immersion levels. In contrast, social presence decreases with increasing immersion. In general, participants performed better in the individual task than in the collaborative task. However, the degree of immersion did not influence the collaborative performance. In addition, we could not determine any adverse effects on depth perception or user experience after use. The present study provides initial contributions to the future of spatial computing in professional settings and highlights the importance of balancing immersion and social interaction in a world where digital and physical spaces seamlessly coexist.

Climate change is one of the most pressing global challenges in the 21st century. Urgent actions favoring the environment’s well-being are essential to mitigate its potentially irreversible consequences. However, the delayed and often distant nature of the effects of sustainable behavior makes it challenging for individuals to connect with the issue personally. Immersive media are an opportunity to introduce innovative feedback mechanisms to highlight the urgency of behavior effects. We introduce a VR carbon calculator that visualizes users’ annual carbon footprint as CO2-filled balloons over multiple periods. In a 2 × 2 design, participants calculated and visualized their carbon footprint numerically or as balloons over one or three years. We found no effect of our visualization but a significant impact of the visualized period on participants’ environmental self-efficacy. These findings emphasize the importance of target-oriented design in VR behavior interventions.

This paper presents an interaction study with a mixed-cultural Intelligent Virtual Agent (IVA) that investigates the impact of perceived ingroup similarity on the empathy of human interlocutors, and the perceived competence and warmth of the IVA. We herefore implemented an interactive scenario with either an IVA that should be perceived as an ingroup member or a neutral IVA. Our results show that the IVA in the ingroup condition triggered significantly more empathy and was rated significantly more competent than the IVA in the neutral condition. Regarding perceived warmth, both IVAs were rated rather high, revealing no significant differences. With these results we contribute to the goal of raising empathy towards IVAs with mixed-cultural backgrounds, to be potentially used in scenarios that aim at studying and reducing implicit racial bias.

This paper presents an interaction study that investigates whether participants adapt their verbal and non-verbal behaviour towards an IVA that shows non-native speech patterns in a way that is congruent to observations previously made in human-human and human-agent interaction. In a between-subjects design, participants either interacted with an IVA speaking native-accented German, an IVA speaking German with a non-native accent, or an IVA speaking grammatically incorrect German with a non-native accent in VR. The results reveal that participants partly adapt their verbal and non-verbal behaviour when interacting with the IVAs with non-native speech patterns, which constitutes important implications for the cost-efficient design of enculturated IVAs.

This paper presents a system consisting of two tools to facilitate the cost-effective implementation of mixed-cultural intelligent virtual agents (IVAs) for both basic and applied research. The first tool automatically introduces grammatical mistakes and non-native accents into a given text to produce synthetic non-native speech. The second tool is a Wizard-of-Oz scenario builder, created with the game engine Unity, that can construct natural interaction scenarios with a dynamic number of (mixed-cultural) IVAs. It enables the control of the appearance and non-verbal behaviour of the IVAs as well as their verbal behaviour by integrating the first tool. Both tools can be either used separately or combined for a wide range of applications in both basic research, for example to investigate the perception of mixed-cultural cues in IVAs, and applied research, for example to create virtual interventions with IVAs.

This paper presents an exploratory study investigating the impact of non-native accented speech on the perception of Intelligent Virtual Agents (IVAs). In an online study, native English speakers watched a video of an IVA holding a monologue whilst speaking English with either a Spanish, Hindi or Mandarin accent that was either recorded by native speakers of that respective language (natural speech) or synthetically generated (synthetic speech). The results showed a significant impact of naturalness of speech on the IVAs perceived warmth and a significant interaction of accent and naturalness of speech on its perceived competence. The naturalness of speech impacted the participants’ perception of the IVA as a non-native speaker of English, and the correctness of the attributed mother tongue in the Spanish and the Mandarin accent condition. These results are a valuable contribution to research on mixed-cultural IVAs in general and non-native speech as a cultural cue more specifically.

This study investigates how different levels of a non-native Turkish accent in German speech are perceived by Turkish-German listeners, using either natural or synthetic speech. The participants listened to six audio recordings and rated the respective speaker regarding her mother tongue, warmth, competence, and intelligibility. The results show that the naturalness of speech had no impact on the non-native speakers’ ability to assign the correct mother tongue to the respective speaker. It did, however have an impact on the speakers’ perceived warmth, competence and intelligibility.

This paper presents an exploratory study which investigates the impact of different non-native accents and the naturalness of speech on the correct assignment of an Intelligent Virtual Agent’s (IVA) mother tongue, as well as its perceived warmth, competence and intelligibility. An online-experiment with a between subjects design was conducted, in which the participants, who were native speakers of German, watched a video of an IVA that spoke German with a non-native accent. The IVA’s speech was either synthetically generated or pre-recorded using non-native speakers. The participants experienced an IVA with either a Turkish, Italian or Polish accent, based on the most frequent accents in the German-speaking area. The results revealed that the IVA’s accent impacted its perceived warmth, but not its perceived competence and intelligibility. The IVA’s naturalness of speech played no role in its classification as a non-native speaker of German but on the correctness of the assigned mother tongue within the Polish accent condition. These results give valuable insight in the perception of non-native speaking IVAs and constitute helpful implications for future research with mixed-cultural IVAs.

This paper presents the implementation of a prototype including a virtual environment in which humans are exposed to a seemingly autonomous Intelligent Virtual Agent (IVA) that ostracises them by using its smartphone during a conversation (a phenomenon called phubbing with adverse outcomes in interhuman communication). Based on the temporal need threat model of ostracism, the effects of phubbing were examined in an online experiment with three conditions (no vs. proactive vs. reactive phubbing) using an IVA. The results showed a significant effect on the warmth attributed to the IVA. Consequently, our study provides partial support that the negative effects of phubbing can be transfered to human-agent interaction. It is thereby not only in line with the underlying theory, but also provides a prototype suitable for future research with phubbing in the context of IVAs.

The present contribution investigates the effects of spoken language varieties, in particular non-standard / regional language compared to standard language (in our study: High German), in social robotics. Based on (media) psychological and sociolinguistic research, we assumed that a robot speaking in regional language (i.e., dialect and regional accent) would be considered less competent compared to the same robot speaking in standard language (H1). Contrarily, we assumed that regional language might enhance perceived social skills and likability of a robot, at least so when taking into account whether and how much the human observers making the evaluations talk in regional language themselves. More precisely, it was assumed that the more the study participants spoke in regional language, the better their ratings of the dialect-speaking robot on social skills and likeability would be (H2). We also investigated whether the robot's gender (male vs. female voice) would have an effect on the ratings (RQ). H1 received full, H2 limited empirical support by the data, while the robot's gender (RQ) turned out to be a mostly negligible factor. Based on our results, practical implications for robots speaking in regional language varieties are suggested.

This paper presents an experiment evaluating the effects of virtual agents' language proficiency on whether they are perceived as a native speakers or not. Our first results indicate that beyond 10% word order mistakes and 25% infinitive mistakes, virtual agents are perceived as non-native speakers, even though their appearance and non-verbal behaviour were not altered. We believe these thresholds constitute interesting guidelines possibly simplifying the design of non-native speaker simulations.

This paper presents an exploratory experiment measuring the role of audio feedback on the illusion of virtual body ownership (IVBO) under non-immersive mixed reality (MR) settings with Human and Non-Human avatars. Our preliminary results revealed that all avatars elicited a similar level of IVBO, despite the addition of audio feedback.