With the rise of generative artificial intelligence, a new partner for interactive, reciprocal, long-term interactions has emerged. Now, it is possible to enter into a new type of relationship, synthetic relationships, i.e. “continuing associations between humans and AI tools that interact with one another wherein the AI tool(s) influence(s) humans’ thoughts, feelings and/or actions” (Starke et al., 2024). Last year, we introduced and defined the term “synthetic relationships” and discussed their opportunities and risks for individuals and society (Starke et al., 2024), particularly in a narrative review on synthetic relationships and social health (Ventura et al., in prep.). For practice, we provided initial policy recommendations (Starke et al., 2024) and proposed a methodological approach for comprehensive research of synthetic relationships (Ventura & Köbis, 2024). Following these initiatives, we are currently taking the first steps towards establishing a synthetic relationships research platform in cooperation with our dear colleagues from the University of Amsterdam.

Our poster provides comprehensive insights into the nascent research field of synthetic relationships - the construct, its relevance for individuals and society, and a proposal for research methodology. This insight will be enriched with current empirical evidence, evidence from our first empirical studies and the empirical studies of colleagues.

Podcasts are widely used to obtain information on important topics, like climate change adaptation and mitigation strategies. While research suggests that individuals tend to listen to podcasts consistent with their preexisting attitudes, individual differences driving such tendencies remain underexplored.

Building on this observation and on theoretical and empirical research from fields like (dis)confirmation biases and intellectual humility, the present work investigated selective exposure – seeking information consistent with one’s attitudes – and selective avoidance – avoiding information inconsistent with one’s attitudes – tendencies when selecting podcast episodes on climate change adaptation and mitigation strategies to listen to. The associations of these tendencies with individuals’ preexisting attitudes, attitude extremity, and intellectual humility were examined by testing the following hypotheses:

H1: On average, individuals select more episodes consistent with their preexisting attitudes than inconsistent ones.

H2: Attitude extremity is positively related to selective exposure and selective avoidance.

H3: Intellectual humility is negatively related to selective exposure and selective avoidance.

H4: Intellectual humility moderates the relations of attitude extremity with selective exposure and selective avoidance, in that higher intellectual humility attenuates these relations.

A quota sample of N = 556 (53% men) from the German adult population participated in an online study. Participants reported their podcast consumption habits and completed validated scales on their attitudes toward climate change adaptation and mitigation strategies and intellectual humility. Additionally, participants were presented with descriptions of six podcast episodes, half of which depicted a negative stance on climate change adaptation and mitigation strategies, whereas the other half depicted a positive stance. Participants were asked to select all episodes they would listen to.

(Generalized) Linear models were used to examine the relations of control variables, participants' attitudes and attitudinal extremity toward climate change adaptation and mitigation strategies, intellectual humility, and their interactions with selective exposure and avoidance scores.

Psychosis and schizophrenia are often portrayed stereotypically in media, including video games. However, “Hellblade: Senua's Sacrifice” and “Goliath: Playing with Reality”, two recent immersive VR games grounded in lived experience, offer a platform to challenge these negative perceptions. By analysing the content of online reviews, this study explores how players describe their experiences with the games and how they perceive depictions of mental illness. A total of 249 reviews were analysed, with 130 for “Hellblade” and 119 for “Goliath”. Game experience references varied across the two, with “Hellblade” reviews predominantly discussing VR affordances (82.7%) and game systems (69.2%), while “Goliath” reviews focused more on narratives (48.7%) and less on VR affordances (21.9%). The majority of references to mental illness were positive, with these primarily found in “Goliath” reviews. The role of immersive VR in reducing stigma, particularly through positive representations of mental illness, is discussed, with implications for game researchers and developers.

Despite legal frameworks such as the Web Content Accessibility Guidelines, individuals with disabilities still face barriers on the internet. This study examines the importance of accessible websites for people with cognitive impairments, aiming to bridge the gap between theoretical accessibility standards and real-world user experience.

The research question adressed in this thesis is: How do selected aspects of digital accessibility impact the user-friendliness for individuals with cognitive impairments? To explore this, a qualitative study was conducted using semi-structured interviews with twelve individuals diagnosed with cognitive disabilities.

Three websites with varying levels of digital accessibility were assessed, both through automated accessibility evaluation tools (WAVE Web Accessibility Evaluation Tool, AChecker) and through user testings. Three key aspects of digital accessibility were examined during the interviews: (1) the availability of plain language, (2) predictable and consistent navigation, and (3) the ability to pause, stop, or hide flashing or moving content. Participants’ perceptions of user-friendliness were evaluated through task-based interactions. The collected data were analyzed using a qualitative content analysis following Kuckartz & Raediker (2022) with the MAXQDA software.

By aligning technical accessibility assessments with user-centered insights, this study contributes to a deeper understanding of how digital accessibility features affect the online experience of individuals with cognitive impairments. The findings offer practical implications for designing more inclusive digital environments.

In a permanently online and permanently connected world (Vorderer et al., 2018), mobile messenger users often experience social pressure to be available to others, that is, the perceived obligation to respond quickly to incoming calls or messages and to regularly check the smartphone for potentially incoming notifications (Bayer et al., 2016; Hall, 2017), thereby challenging users’ mental health in many ways: On the one hand, experiencing availability pressure elicited negative affective states (Van Gaeveren et al., 2024), and giving in to availability pressure led to negative perceptions such as self-control failure (Halfmann et al., 2021). On the other hand, experiencing availability pressure was positively related to relationship quality (Li & Chan, 2024) as well as competence and relatedness satisfaction (Reinecke et al., 2014), and giving in to availability pressure could provide a sense of social support (Utz, 2020) and belonging (Hall et al., 2023). Given these ambivalent effects, availability pressure can be linked to both well-being and ill-being, representing two aspects of mental health (Meier & Reinecke, 2021).

The present submission aims at clarifying these heterogeneous findings by investigating how appraisal processes and individual resources affect the relationship between availability pressure and users’ well-/ill-being (operationalized as affect, vitality, and stress). Applying the Integrative Model of Mobile Media Use and Need Experiences (IM3UNE; Schneider et al., 2022), it conceptualizes the satisfaction versus frustration of autonomy, competence, and relatedness needs as mediators and the health-promoting traits mindfulness, self-control, meaning in life, and sense of coherence as moderators.

In a cross-sectional pilot study (N = 400), we first explored the relationship between these traits. Afterward, we conducted four pre-registered online experiments (N1 = 265, N2 = 297, N3 = 193, N4 = 162), manipulating availability pressure to test the IM³UNE paths described above. Following open science guidelines, each study’s material can be found on OSF (https://t1p.de/bv5xj).

As followbots become increasingly integrated into workplaces and social environments (Almokdad & Lee, 2024), understanding their impact on human perceptions of leadership, trust, and social presence is crucial. This study examines how followbot autonomy (autonomous vs. other-controlled), leader gender (male vs. female), and followbot functionality (malfunctions vs. does not malfunction) influence impressions of leadership quality, trust, and social presence. Using Social Presence Theory (SPT) (Short, Williams, & Christie, 1976) and the Computers Are Social Actors (CASA) paradigm (Nass, Steuer, & Tauber, 1994), we explore how followbots function as social cues shaping perceptions of leadership. SPT suggests that technology’s perceived ability to facilitate meaningful interaction affects interpersonal judgments, while CASA posits that people apply social rules to machines, influencing how followbots factor into leadership evaluations.

This study employs a 2 × 2 × 2 experimental design in which participants observe a leader working with a followbot under one of eight conditions and then assess the leader’s effectiveness, trustworthiness, and social presence. The following research questions guide our investigation:

(1) How do followbot autonomy and malfunctioning behavior influence perceptions of leadership quality, trust, and social presence?

(2) How does leader gender affect perceptions of leadership quality, trust, and social presence in human-followbot collaboration?

(3) Does a malfunctioning followbot decrease perceptions of a leader’s trustworthiness and social presence compared to a properly functioning followbot?

Findings will be discussed in light of these frameworks to provide insights into human-followbot collaboration. This research has implications for designing and deploying robotic assistants in leadership contexts, emphasizing the importance of reliability, perceived control, and gender dynamics in fostering trust and effective teamwork in human-robot interactions. Additionally, this study contributes to the broader discourse on human-technology interaction by highlighting how automated agents mediate social relationships, workplace hierarchies, and organizational dynamics, ultimately shaping future leadership paradigms.

Whereas a substantial amount of empirical research on misinformation exists, one phenomenon with increasing societal awareness has largely been disregarded: the liar’s dividend, i.e., the leveraging of uncertainty due to the existence of misinformation to one’s own advantage. So far, only one series of studies examined whether politicians may escape accountability of making insensitive statements by discrediting genuine evidence as fake (Schiff et al., 2024): Whereas for textual evidence, politicians claiming the personal scandal to be due to disinformation gained higher support compared to giving no response or apologizing, the results for audiovisual evidence were inconclusive. Yet, concluding that deepfake attributions are largely ineffective for audiovisual material (i.e., do not pay a liar’s dividend) may be premature. While rapid technological advances render deepfakes increasingly realistic and difficult to distinguish from genuine content, public debates regarding deepfakes within the context of political conflicts may have increased the credibility of deepfake attributions. Therefore, the goal of the current study was to re-examine the liar’s dividend for audiovisual material. To this end, we conducted an online experiment presenting participants (N = 182) with genuine video evidence of a scandalous political incident followed by a supposed statement of the respective politician including either false deepfake claims or an apology. We examined whether the deepfake claim compared to an apology does pay a liar’s dividend in terms of perceived leadership ability, increases the likelihood that the video evidence is misidentified as a deepfake, and decreases general trust in media. We investigated further whether misidentification of the evidence as a deepfake mediates the liar’s dividend effect of deepfake claims and is negatively associated with individuals’ trust in media. Results are interpreted and discussed against prior evidence. Conclusions on the impact of false deepfake claims as well as future research directions are presented.

Theoretical Background. Phubbing – the temporary disregard of an interaction partner in favor of one’s smartphone – is a common phenomenon in today’s digitalized world, often linked to negative effects on mental well-being and interpersonal relationships. However, the effects of phubbing are not universally negative—while some individuals experience detrimental consequences, others remain unaffected. This inconsistency highlights the need to identify for whom phubbing is particularly harmful. Building on recent findings, this study integrates the Temporal Need Threat Model of Ostracism with Attribution Theory to examine how attributional styles (internal vs. external) influence the effect of phubbing on emotions, self-esteem and the evaluation of the interlocutor.

Research question. We assume that the effects of phubbing on emotional states as well as the self-evaluation and evaluation of the interlocutor vary depending on the attributional style. More specifically, we hypothesize that attributional style moderates the effect of phubbing on guilt and anger, such that internal attribution increases guilt, while external attribution increases anger. Secondly, it is assumed that the attributional style moderates the effect of phubbing on self-esteem and interlocutor’s warmth and politeness, such that internal attribution decreases self-esteem, while external attribution leads to more negative perceptions of interlocutor’s warmth and politeness.

Methodological approach. A VR-based human-agent interaction using a Wizard-of-Oz scenario is implemented, in which participants converse with a virtual person in a VR-simulated café. Participants (planned N = 152) are randomly assigned to a phubbing (experimental condition = EC) or a no-phubbing condition (control condition = CC). In the EC, the virtual person uses a smartphone during the conversation with the participant. Before the VR interaction, attributional style is assessed, while guilt, anger, self-esteem, and the evaluation of the virtual person’s warmth and politeness are measured afterwards. Results are analyzed with regression analyses.

Narratives shape the perception of fictional characters. This qualitative-focused mixed-methods study investigates the impact of narratives in video games on player avatar customization. In two studies participants were presented with both a subtly stereotypical and sexist narrative as well as a neutral story about a young student’s first weeks at university; study 1 investigated the effect of the narratives on the creation of female avatars and study 2 extended the design to male avatars. Importantly, participants were not made aware of the subtle differences between the narratives, nor where they able to pinpoint the differences when asked about it. Based on these narratives, but oblivious to their sexist connotations, participants were instructed to create an avatar reflecting the character in the story using the game The Sims 4 by Electronic Arts. During the avatar creation they were asked to think aloud and their thoughts were recorded and analysed. Sexist beliefs intrinsic to the person may influence the creation of the avatar, but more importantly the narratives themselves can change how people form a picture about a character in their mind. Beliefs and narratives together can create the notion that one gender is superior to the other or influence how a person perceives their environment. Based on these beliefs harmful stereotypes can form about gender (e.g., the belief that women are weak and are there to serve the man who protects them, or that women try to control men sexually), which may lead to discrimination. Thus this study highlights the influence of narratives on shaping player perceptions and choices. Overall, this research contributes to understanding how sexist beliefs manifest in virtual environments and provides a foundation for future studies to explore interventions that promote more inclusive and balanced video games.

Online tools provide users self-presentational opportunities, both deliberate and unintended. Two theories relevant to online self-presentation effects on self-concept have recently gained prominence, and yet require a direct test to contrast their predictions: Proteus effect (Yee & Bailenson, 2007) and identity shift theory (Carr et al., 2021) and the Proteus effect. Both suggest self-presentation in an online setting can trigger a corresponding change in self-perception, but differ regarding the mechanism for change.

Proteus effect refers to the process by which an individual adjusts their behavior to conform to a digital self-representation (Yee & Bailenson, 2007), focusing on the effects of visual self-representation. Alternately, identity shift refers to the process by which an individual adjusts their attitudes and beliefs to conform to a digital self-presentation (Gonzales & Hancock, 2008), focusing on the effects of textual self-presentation. Both theories predict changes in self-concept; but based on disparate processes that can often co-occur online, particularly via videoconferencing. Hypotheses predict main effects of both visual self-representation and textual self-presentation (H2), as well as an interaction effect whereby textual self-presentation has a stronger effect than visual self-presentation (H3) due to expectation that behavioral confirmation is stronger when with self-expectations (i.e., identity shift) rather than others expectations (i.e., Proteus effect) (Bem, 1972).

N = 120 students (≈30 per each of the four conditions) at a mid-sized Midwest American university will take part in a lab study about how people self-present in videoconferencing tools like Zoom. Participants will be randomly assigned to (a) textually self-present themselves as either morally good or bad while concurrently (b) seeing themselves with an augmented reality (AR) filter depicting them with either a halo (morally good) or bandit mask (morally bad). Difference in post-treatment self-assessments of participants’ morality will be used to test hypotheses using contrast analyses.

The term online vigilance (OV) describes “users’ permanent cognitive orientation towards online content and communication as well as their disposition to exploit these options constantly” (Reinecke et al., 2018, p. 1/31). Though OV should not be equated with pathological behavior, such as Internet addiction (Reinecke et al., 2018), a strong orientation towards the online world could carry certain risks, like increased stress (Freytag et al., 2021). However, some recent research indicates that OV “does not pose a serious threat to affective well-being in everyday life” (Johannes et al., 2021, p. 581) and that it can also be beneficial in terms of relatedness need satisfaction (Reinecke et al., 2018).

The current study examines the relationship between OV and 3 indicators of psychological well-being (stress, depression/anxiety, and loneliness). We hypothesize that OV can be beneficial as well as detrimental for well-being (Reinecke et al., 2018) and that this relationship is mediated by information overload (IO; Misra & Stokols, 2012) and perceived online social support (OSS; Frison & Eggermont, 2016). On the one hand, we assume that OV can elicit an overly online engagement and that this might lead to IO, which, in turn, can have negative impacts on well-being (e.g., Matthes et al., 2020). On the other hand, OV and the associated greater use of online media could also be associated with more perceived OSS (Reinecke et al., 2018), which, in turn, could have positive impacts on well-being.

The study was conducted as a stratified (criteria: age, gender, region) online survey among 1,075 German Internet users (mean age: 44.93 years; 525 females, 549 males, 1 diverse). All variables were measured using established scales. To test the hypothesized relationships, we performed structural equation modeling using IBM SPSS Amos. Indirect paths were tested with bootstrapping (5,00 samples, 95% bias-corrected confidence intervals).

Climate communication often struggles to engage diverse audiences due to ideological polarization and defensive reactions, particularly in response to fear-based messaging (Huddy, Davies, & Sandor, 2020). This study explores how cognitive reappraisal, an emotion regulation strategy that reframes distressing narratives into constructive and empowering messages (Gross, 1998; Gross & John, 2003), can improve the effectiveness of climate communication. Additionally, it examines how virtual reality (VR) interventions can enhance cognitive reappraisal by immersing users in experiential scenarios that reduce ideological resistance and foster emotional engagement (Kitson et al., 2024).

Using insights from media psychology, emotion regulation theory, and digital persuasion research, this study proposes a two-phase experimental design:

Phase 1: Emotional and Cognitive Impact

Participants (N=400), stratified by political ideology, engage with three messaging conditions:

Reappraisal-Based Messaging: Frames climate action as achievable and collective.

Fear-Based Messaging: Highlights catastrophic consequences.

Neutral Messaging: Presents factual information without emotional framing.

Measures: Emotional responses (hope, fear, self-efficacy), ideological defensiveness, and policy support (Gross & John, 2003).

Phase 2: Behavioral Intentions and VR Integration

Participants (N=600) interact with VR experiences designed to elicit cognitive reappraisal (Kitson et al., 2024).

VR scenarios include perspective-taking (experiencing climate change from multiple viewpoints), embodied interaction (exploring sustainable solutions in virtual environments), and narrative framing (personalized storytelling to emphasize agency and collective action).

Measures: Behavioral engagement in sustainable actions (Panno et al., 2020).

This research contributes to media psychology and digital behavior change by demonstrating how cognitive reappraisal and VR can reshape climate discourse. Findings will provide actionable insights for environmental communicators, media strategists, and policymakers seeking to design inclusive and persuasive climate campaigns that bridge ideological divides.

Theoretical background.

Beginning the day with social media use (SMU) is a common habit that influences eudaimonic (EWB) and hedonic well-being (HWB). Prior research on SMU has predominantly examined interventions aimed at reducing or completely abstaining from SMU; however, evidence suggests these approaches may not yield the anticipated benefits for well-being. Expanding on these findings, our study explores whether substituting SMU with a meaningful activity has a differential impact on EWB and HWB over time.

Research questions.

(1) How does the intervention impact morning EWB and morning HWB at the end of the intervention compared to baseline levels? (2) How do morning EWB and morning HWB develop over time during the intervention? (3) How do EWB and HWB change from morning to evening and how do these changes differ between the intervention groups?

Methodological approach.

We conduct an intensive longitudinal randomized controlled trial (RCT)

with N = 200 participants, randomly assigned to either (IG1) reducing morning SMU or (IG2) reducing and substituting SMU with a meaningful activity. Over 15 working days, participants complete daily morning and evening assessments of EWB, HWB, SMU, smartphone use and mindless scrolling using self-reports and iOS screen time. We hypothesize that morning EWB will increase, with a stronger and more sustained effect in IG2. Conversely, morning HWB may initially decline but is expected to return to baseline by the end of the intervention. Linear mixed-effects models analyze within- and between-group changes. This study will provide insights into how EWB and HWB evolve throughout the intervention, refining strategies for healthier morning routines and digital well-being. The findings will contribute to developing more effective interventions that go beyond simple SMU reduction, emphasizing the role of meaningful substitution.

Robots can benefit children’s musical instrument practice as social companions (Song et al., 2021). However, the integration depends significantly on the attitudes and involvement of key stakeholders (e.g., parents and teachers). Studies indicated that parental attitudes shape children's acceptance of robots (Oros et al., 2014), while teachers’ perceptions influence technology adoption in education (Blackwell et al., 2013). Despite their decision-making power, stakeholders were often excluded from the interaction loop between children and robots. Prior research emphasized the importance of stakeholder engagement (Song et al., 2022), aligning with participatory design approaches (Good & Robertson, 2006). In classroom settings, interventions like Cellulo have granted teachers control over robot-assisted learning (Özgür et al., 2017), yet informal settings like music practice remain under-explored. Our research aimed to involve parents and teachers in shaping children’s musical practice with social robots, enhancing both engagement and effectiveness.

The objective of this research is to answer the following research question: What are the distinct needs and expectations of parents and teachers regarding social robots in children's musical instrument practice, and how can these insights inform the design of stakeholder-inclusive robot functionalities?

The current study employed a two-phase approach: stakeholder interviews and co-design sessions. First, semi-structured interviews were conducted with 4 parents and 4 music teachers to explore their distinct needs and expectations regarding social robots in children's musical practice. The interview guidelines covered topics such as practice frequency, supervision needs, preferred robot functionalities, information presentation, and personalization. Insights from interviews were used to inform the next phase. In the second phase, four co-design sessions were held separately with parents and teachers to collaboratively develop solutions. These sessions incorporated participatory design methods (Zamenopoulos & Alexiou, 2018), encouraging stakeholders to connect past experiences with future possibilities. Activities included knowledge probing, critical reflection, idea generation, future envisioning, and prototyping.

Artificial intelligence (AI) is increasingly integrated into decision-making across various domains, raising concerns about both over-reliance and under-reliance on AI-generated advice. Appropriate reliance, defined as the ability to accept correct AI recommendations while rejecting incorrect ones (Schemmer et al., 2022), is crucial for optimizing human-AI collaboration. While previous research has examined factors influencing AI reliance individually, there is a lack of comprehensive models testing how expertise, trust, self-confidence, and task difficulty interact to shape reliance behaviors. Parasuraman and Riley (1997) identified these factors as key determinants of reliance on automated systems, a framework later supported by Goddard et al. (2012), who highlighted the need for research that systematically examines their interrelationships.

This study addresses this gap by testing an integrated model of AI reliance, examining how expertise influences trust, self-confidence, and reliance behaviors in a controlled setting. Using chess proficiency as a domain-specific proxy for expertise, participants completed a series of chess puzzles first independently and then with AI-generated recommendations. Structural equation modeling was employed to analyze the relationships between expertise, trust, self-confidence, and reliance behaviors, providing a holistic view of how users calibrate their reliance on AI decision support systems.

The study seeks to answer two primary research questions: (1) What factors contribute to under-reliance among experts and over-reliance among non-experts in AI decision support? (2) How does expertise influence trust, self-confidence, task difficulty, and appropriate reliance in AI-assisted decision-making? By integrating multiple psychological determinants of reliance into a single model, this research aims to advance theoretical understanding and inform the design of AI systems that better support user decision-making across domains. Findings from this study have the potential to contribute to the development of AI tools that optimize reliance by accounting for individual differences in expertise and decision-making strategies.

Humans may soon need to collaborate with robots in the workplace. For a smooth integration, it is essential that employees accept robots as co-workers. This raises the question of how to foster robot acceptance in collaborative tasks. One seemingly simple approach is to increase their likability. Research in social psychology suggests that people tend to like others more when they share similarities, the so-called ‘Similarity Effect.’ Studies show that similarity leads to more favorable evaluations and a greater willingness to cooperate (Byrne, 1961, 1971; Harrison et al., 1998; van Emmerik & Brenninkmeijer, 2009). According to the Media Equation Theory, such phenomena can also be transferred to the interaction with robots (Nass & Moon, 2000; Reeves & Nass, 1996). Similarity can be established on either a personality or an attitudinal level. Regarding personality similarity, research with robots yielded mixed results, whereof some support the effect, while others suggest a complementarity attraction effect (e.g., Isbister & Nass, 2000; Kuchenbrandt et al., 2013; Lee et al., 2006; Park et al., 2012; Tapus & Mataric, 2008). In contrast, studies on attitudinal similarity indicate that aligning a robot’s attitudes with those of humans leads to more positive evaluations (Bernier & Scassellati, 2010). However, no studies have yet examined the impact of attitudinal similarity on cooperation with robots in the workplace. To address this, we conducted a between-subjects laboratory experiment with N = 55 participants, who collaborated with a humanoid robot on a brick-scanning task. Before the collaboration began, the robot either expressed the same or a different music preference as the participant. We hypothesized that participants in the similarity condition would perceive the robot as more likable and be more willing to continue assisting it both in the moment and in the future measured via self-report.

Prior research has emphasized that experiences of social exclusion (e.g., being ignored by peers or blocked from online discussions) threaten several fundamental needs and lead to a severe decline in people’s emotional well-being. More specifically, it has been found that exclusion episodes negatively impact a person’s needs for belongingness, control, self-esteem, and meaningful existence, while also worsening the immediate mood state. In response to this worrisome pattern, scholars have investigated the restorative power of media—suggesting that certain formats and types of content might provide socially excluded individuals with beneficial coping opportunities. Yet, whereas empirical findings clearly underscore socially relevant coping effects by traditional 2D media such as television, much less is known about the potential well-being and need repair offered by more immersive media environments such as virtual reality (VR). Addressing this research gap, our pre-registered laboratory experiment (see https://aspredicted.org/vk52-ck8p.pdf) will examine the effects of a VR-based social intervention to restore individuals’ needs and improve their affective state after an exclusion episode. First, participants will be exposed to a well-established task to induce social ostracism, the so-called Cyberball game; the success of this induction will be checked by comparing their self-reported inclusionary status before and after playing. Following a first measurement of emotional well-being and threatened needs, participants will be randomly assigned to experience a (pre-scripted) social situation either in a VR headset or on a 2D monitor. After our intervention, outcome variables will be measured again, yielding pre-post difference scores. It is assumed that VR not only leads to a stronger improvement of emotional well-being and need restoration than the 2D version, but also that this advantage will be mediated by certain perceptual variables uniquely connected to the VR experience—namely, participants’ social presence, spatial presence, and self-presence during the intervention.

Social robots are increasingly present in public spaces (Gasteiger et al., 2021), offering innovative services while simultaneously raising privacy and security concerns due to unclear data handling practices (Lutz et al., 2019). A lack of transparency in data collection and processing can lead to uncertainty and mistrust among humans, as well as misunderstandings regarding the scope of the robot’s functionality (Fischer et al., 2018). Conversely, ensuring a high level of transparency fosters greater trust in robots (Sanders et al., 2014) and is crucial from an ethical standpoint (McBride, 2020). While extensive research has been conducted on creating explainable content, little attention has been given to how data transparency—addressing uncertainty and concerns about the handling of personal data—is effectively communicated to users (Schött et al., 2023).

Based on previous transparency approaches, we conducted a two-part laboratory experiment with 106 participants who interacted with the social robot Pepper in a virtual reality (VR) environment and explored, in addition to previous unimodal approaches, multimodal approaches and compared them with a control group without transparency declarations. Participants were randomly assigned to one of six conditions: no transparency declaration, robot speech and tablet (multimodal), robot speech and external posters (multimodal), external posters (unimodal), tablet (unimodal), and speech (unimodal). Tasked with obtaining a library card with Pepper’s assistance, participants had to provide personal data to the robot, making them explicitly aware that their data was being actively collected and processed. To assess the effectiveness of these transparency declarations, we measured participants' understanding of the transparency statements, trust in the robot (Jian et al., 2000), technology acceptance (Venkatesh & Bala, 2008), perception of the robot (Spatola et al., 2021), and anxiety toward robots (Nomura et al., 2006). The findings contribute to a better understanding of how robots and the deploying institution are perceived under various transparency declarations.

The capacity to imagine alternatives to the status quo plays a crucial role in motivating collective efforts for change (Wright et al., 2022). This study examines how writing narrative fiction may promote the imagination of a sustainable future and foster creative self-efficacy. Fictional stories, particularly science fiction, can reduce the perceived threat of societal transformations and improve the imagination of alternative futures (Appel et al., 2016). Fiction writing may be especially effective due to the deep engagement it requires. Research suggests that actively contributing to a narrative can enhance its persuasive impact (Lillie et al., 2023). Engaging in creative activities may also promote well-being by fostering a sense of accomplishment and efficacy (Forgeard & Eichner, 2014), yet the effects of fiction writing remain underexplored.

However, some people may not possess sufficient writing skills or knowledge about a complex issue like climate change to benefit from writing climate fiction. Those individuals may be supported by co-creating the story using generative AI such as ChatGPT (Doshi & Hauser, 2024), which may help to optimize the task challenge. However, reduced cognitive effort may also weaken the positive effects of creative writing. This study investigates how the effects of writing climate fiction may change by using ChatGPT as a collaborative partner, considering individual differences in creative writing competence and subjective climate knowledge.

We present a two-group lab experiment (between-subjects factor: writing instruction) with a pre-post measurement of the dependent variables (environmental cognitive alternatives, creative self-efficacy). Participants (N = 144) wrote a climate fiction story either independently or collaboratively with ChatGPT-4. The writing prompt asked them to imagine effects of climate change on their home region, a desirable future 50 years from now, and describe how it was achieved.

For geographically dispersed groups of friends, group chats are an efficient way to stay in touch. If and how these chats can contribute to connectedness over distance was investigated in our studies.

Group chats may also have negative effects. FOMO – the fear of missing out – is typically associated with SNS. Group chats with close friends should be less likely to cause FOMO, as FOMO stems from an unmet need to belong. However, in a geographically dispersed group, the situation is different: Messages about a group activity that one member has missed due to the distance can be perceived as exclusionary and can elicit FOMO.

The severity of the exclusion experience varies depending on who is seen as responsible for the exclusion and how intentional it is perceived to be. Therefore, in a dispersed group, perceived exclusion and FOMO may depend on who moved away and what intentions are attributed to the messages.

To avoid FOMO, it's crucial to know what characteristics of messages promote it. As one of the psychological processes underlying FOMO is social comparison, we investigated whether chat messages depicting an individual’s rewarding experiences can induce FOMO.

Research Questions

Do connectedness and FOMO differ between dispersed and geographically close groups of friends?

How do message characteristics influence FOMO?

To what extent do messages about missed group activities or individuals’ activities elicit FOMO?

Two preregistered studies measured perceptions of group chats:

A survey assessed the characteristics of messages of participants’ own group chats and the emotions associated with them.

In a vignette experiment with two conditions, participants were asked to imagine receiving messages about a group activity that they had missed - either because they or other group members had moved away. The experimental conditions were compared in terms of connectedness, FOMO and attributions of messages.

Media multitasking has become an increasingly prevalent behavior among students in the digital age, raising essential questions about its impact on learning outcomes. This practice, defined as the simultaneous engagement in multiple activities with at least one involving media (e.g., Beuckels et al., 2021), has been linked to cognitive challenges, such as diminished attentional control and academic performance (Lau, 2017; van der Schuur et al., 2020). As digital media becomes more ubiquitous in educational contexts, understanding the nuanced effects of multitasking on students' learning experiences becomes increasingly critical.

Recently, study-with-me videos (SWMs) have become a global media phenomenon (Lee et al., 2021). These videos depict calm, aesthetically pleasing study environments, providing virtual companionship during learning sessions (Ren et al., 2025). Students intentionally engage with SWMs while studying, creating a media multitasking context that challenges traditional perspectives. Established multitasking research, grounded in limited-capacity theories (e.g., Lang & Chrzan, 2015), suggests that cognitive resources are finite, and their division across tasks typically leads to performance declines. However, emerging pedagogical perspectives propose that SWMs may enhance learning motivation and performance (Ren et al., 2025). This study aims to reconcile these competing viewpoints by examining whether SWMs can positively influence task-induced motivation (RQ1) and cognitive performance (RQ2) in student populations.

We therefore implemented a 2x1 between-subjects experimental design with 111 participants (aged 18-30). All participants completed measures assessing multitasking tendencies, polychronicity, and learning motivation. Cognitive performance was evaluated using an operation span task, which tested recall and recognition accuracy. Participants were randomly assigned to the experimental conditions (combination of task and SWM stimulus vs. no SWM stimulus).

This study contributes to the broader discourse on media multitasking by providing empirical evidence regarding SWMs' cognitive and motivational effects. It additionally offers practical insights for educational contexts seeking to leverage digital tools in learning environments.

In the context of professional environments, robotic systems are anticipated to increasingly shape work distribution and influence social dynamics. A pivotal question arises regarding the (social) roles that robots will adopt and how these roles will be perceived by humans. This study investigates whether the assessment of a robot is influenced more by its product characteristics or by its perceived social categorization.

To explore this, we conduct a between-subjects online study in which we manipulate the robot's categorization based on nationality (German, Japanese, Turkish) and its presentation style (as a product versus as a companion) in the context of patient care assistance. German participants (n=216) watch a video featuring Pepper robot and evaluate its perceived social attributes, design, and contact intentions. Additionally, we assess ingroup identification with Germans, general attitudes and stereotypes towards the respective countries and their products, as well as social dominance orientation.

Through this research, we aim to shed light on two competing explanations for the anticipated variations in robot evaluations: (1) the influence of their social categories, particularly regarding whether they are perceived as part of the ingroup or outgroup. We further predict that the difference in evaluation will be more pronounced when robots are framed as companions. The second explanation focuses on (2) the product attributes of the robots, which may be affected by perceptions related to their "Made in" labels. Regarding this, we hypothesize that attitudes and stereotypes towards countries and products will be positively related to the evaluation of the robots, regardless of ingroup status. We expect to find a correlation between these attitudes and assessments of the robots' social attributes, contact intentions, and design preferences.

The reasons why robots may be perceived differently can inform technology development and marketing and give guidance for successful deployment of robots in companies and work teams.

In today’s digital age, the growing popularity of online social platforms and engagement in virtual environments has provided individuals with more opportunities to explore and express themselves in novel ways (Nagy & Koles, 2014; Hussain & Griffiths, 2008). In virtual spaces like video games, players can embody avatars, facilitating self-representation and self-exploration (Chan, 2022). These avatars often reflect idealized versions of the self, offering insights into how individuals explore their identities in virtual environments (Sibilla & Mancini, 2018).

The present study investigates the impact of avatar-based self-representation in offline video games on mood and the influence of self-esteem, body-esteem, and self-exploration, drawing on Higgins' self-discrepancy theory (1987). Specifically, we examined whether discrepancies between ideal and actual self-representations influence emotional and self-perceptual outcomes. We also explored how perceptions of the virtual world, whether as separate from or an extension of reality, shape self-exploration through avatars.

A total of 69 participants (53.6% male, 46.4% female; Mage = 24.7, SD = 5.53) took part in a lab study, during which participants created avatars over 10 minutes in the video game Fallout 4 and subsequently played the game for another 10 minutes. Avatar-Self-Discrepancy and Avatar-Self-Coherence items (Villani et al., 2016) were used to measure how closely avatars aligned with participants' real- or ideal-selves. Additionally, participants reported their motivations for avatar design. To assess mood, the Positive and Negative Affect Schedule (PANAS, Watson et al., 1988) was administered pre- and post-gameplay. Additionally, participants rated measures of self-esteem (Rosenberg Self-Esteem Scale, Rosenberg, 1965) and body-esteem (Body-Esteem Scale for Adolescents and Adults, Mendelson et al., 2001) before gameplay.

This study contributes to understanding how self-discrepancy in avatar creation may affect emotional and identity-related processes, highlighting the psychological impacts of virtual self-representation.

Game player relationships have traditionally been examined through parasocial interaction frameworks, particularly in otome games where players develop romantic connections with male non-player characters (NPCs) (e.g. Kreissl et al., 2021; Song & Fox, 2016). However, this perspective fails to capture the reciprocal nature of player-NPC relationships (Scriven, 2023), where mutual exchange and interdependence are common—that is, exchanges are social (two-way) rather than parasocial (one-way). Such relationships are especially relevant in otome games, which are designed as romantic dating simulations, require players to see other characters as authentic social others.

In otome games, the core gameplay mechanic revolves around building and maintaining relationships with NPCs through ongoing exchanges (e.g. romantic dialogue and fight along experience), which is especially relevant through the lens of social exchange theory (SET; Blau, 1964). Likewise, gaming research has shown that enjoyment stems from the satisfaction of basic psychological needs: autonomy, competence, and relatedness (SDT; Ryan & Deci, 2000, Tamborini et al., 2011). While relatedness is the most obvious connection to reciprocity in otome games, we also explore how perceived reciprocity might influence autonomy and competence. Our study combines SET and SDT to examine how perceived reciprocity influences player enjoyment in otome games through need satisfaction.

We will recruit participants from online otome gaming communities. Participants will complete an online survey assessing: Perceived reciprocity, measured using an adapted version of the Traupmann-Utne-Walster (TUW) Scales (Traupmann et al., 1981); Need Satisfaction, measured by Player Experience of Need Satisfaction scale (Ryan et al., 2006; Tamborini et al., 2011); Enjoyment, measured using entertainment scale (Oliver et al., 2016; modified by Bowman et al. 2024). An exploratory mediation model will be examined using Hayes’ (2022) PROCESS macro, model 4 for parallel mediators.

According to Media Equation Theory, people react to technical devices in a similar way as to human counterparts (Reeves & Nass, 1996), and social theories can be applied to interactions with technologies such as robots (Riether et al., 2012). One of these theories is the Social Inhibition Effect, according to which the mere presence of others can lead to a decline in performance when performing difficult tasks (Zajonc, 1965). While previous research has also confirmed this effect for human-robot interactions (Wechsung et al., 2014), the influence of previous experience with robots has been neglected. However, previous experience with robots can have a lasting effect on our perception and attitude towards them, strengthening trust and increasing sympathy (Sanders et al., 2017). Building on this, we questioned whether the influence of the presence of a humanoid robot affects human performance in difficult tasks similar to the presence of another human, and if previous experiences with the robot alter the impact. To test this research question, we conducted an experimental study with N = 45 subjects that were randomly assigned to three groups: observation by a robot without prior experience, observation by robot with prior experience, and observation by a human. The experimental sessions were remotely conducted via Zoom, where subjects solved 20 anagrams in two rounds. In the first round, they solved the anagrams unobserved with their camera turned off. In the second round, they were observed by human or a robot via camera. In the condition with previous experience, participants were introduced to the robot before they had to solve the task in its presence. The findings of this work contribute to our understanding of the factors influencing the experience of interactions with robot and provide important starting points for the future design of social robotic systems.

The perception of novel technologies and society's acceptance play a crucial role in the innovation process. While established models often assume that decision-making processes are primarily based on a rational assessment of benefits and risks, more recent approaches show that psychological mechanisms also play a significant role. These mechanisms can influence decisions that go beyond rational considerations. Due to its limited public awareness, terahertz technology seems particularly promising for investigating the impact of psychological factors on the acceptance of new technologies. Since the majority of society has little prior experience with terahertz technology, it allows for the capture and analysis of unbiased perspectives and the development of opinions about innovative technologies.

Our investigation explores, through exploratory semi-structured interviews (N = 17), how people generally perceive novel technologies, to what extent they use them, and what advantages and disadvantages they associate with them. Additionally, we analyze which technologies are perceived as particularly risky, where skepticism exists, and which innovations evoke excitement. Another focus is on the form and degree of trust placed in various technologies. Subsequently, the same approach was applied to terahertz technology, which was presented through specific applications: body scanners at airports, skin cancer screenings, and microflyers for detecting viruses and bacteria. Using these examples, we examine how participants are generally inclined to view a completely unfamiliar technology, under what conditions they would support its use, and to what extent specific applications evoke acceptance or rejection.

The study provides insights into how people perceive and evaluate novel technologies. It particularly highlights the role of trust, skepticism, perceived risks, and information sources. Furthermore, it sheds light on whether and to what extent unfamiliar technologies are evaluated when only limited information is available. These findings contribute to the development of strategies for the effective introduction and communication of new technologies.

Virtual Reality (VR) exceeds common social media platforms by its affordances that allow users to embody either themselves or fictional characters and meet in virtual spaces to interact with each other. Unfortunately, Sexual Harassment (SH) is a significant and common risk for the 171 million VR users worldwide each year. There are multiple intervention measures, for example blocking or muting harassers and entering a safe zone, to reduce the number of affected people. Additionally, users function as volunteer moderators and try to intervene in SH incidents. However, a lack of resources either for the implementation and the moderation itself limits the potential efficiency as an intervention measure. In recent years, scholars proposed Artificial Intelligence (AI) based moderation, where the system could filter written or spoken language to detect harassing speech and take immediate action. Given the constant rising numbers of SH incidents, is it necessary to effectively support harassed individuals and ensure their safety. To do so, it is crucial to adopt the victims’ perspective and assess their views on AI-based moderation and its perceived potential as an adequate intervention measure. Furthermore, evaluation of the efficiency of moderation includes understanding the role of a human vs. an AI-moderator and reasonings why they can be perceived as helpful for the victim in intense social situations. Therefore, a qualitative interview study with six to eight participants that experienced SH in VR before will be conducted, recruited via reddit and discord. The study aims at gaining insights into knowledge and opinions, regarding human vs. and AI-based moderation of those who experienced SH. Further, it will enlighten the (perceived) role and importance of the moderator of both entities to formulate suggestions about its effective implementation to decisively decrease the amount of affected people of SH in VR.

Gratitude plays a key role in enhancing life satisfaction and reducing psychological distress. This project explores the potential of large language models (LLM) to support gratitude practices through their integration into gratitude journaling routines.

While traditional gratitude journals—structured around reflective prompts—have proven effective in improving mental well-being (Davis et al., 2016; Bohlmeijer et al., 2021), they often suffer from low user engagement because they can feel repetitive or obligatory. As gratitude is experienced more intensely during social interactions (Algoe, 2012), scientific studies (Lee et al., 2023) have utilized chatbots to increase gratitude. These studies have relied on explicit expressions by the system rather than on dynamic interactions with a LLM during a personal gratitude routine. To advance research in this area, an LLM-supported gratitude journal was developed providing dynamic feedback.

Two trials were conducted. In the first trial, participants completed a single journal entry with measurements taken immediately before and after the entry. This trial compared two conditions—the LLM-supported condition and a control—by examining changes utilizing PANAS, GQ-6 and warmth and competence.

In the second trial, participants maintained a journal for seven consecutive days and were assigned to one of three conditions: 1. traditional journal (control), 2. journal with LLM-support, 3. journal with LLM-support and previous entries included in the initial prompt. The third condition was added to test whether enhanced contextualization and personalization of the responses affect the measurements. Surveys were administered at baseline, mid-trial, and post-trial, measuring PANAS-SF, GQ-6, IOS, RoSaS and BPN-TU.

This study examines the effects of different levels of LLM-supported journaling on gratitude expression, wellbeing and the formation of a bond between users and the system. The insights gained from this study can serve as a basis for future research on LLM support for gratitude practices and their long-term effects on well-being.

As artificial intelligence and robotics continue to evolve, examining how robots influence human decisions becomes more relevant. Especially in the context of moral decisions, the interplay of humans and AI systems and their embodiment gains importance (Glück, 2022; Rau et al., 2013, Straßmann et al., 2020). Previous studies have demonstrated that people evaluate moral decisions differently depending on whether they are made by humans or AI systems (Chugunova & Luhan, 2022; Zhang et al., 2023). Research indicates that individuals tend to delegate responsibility to AI systems, particularly in complex or emotionally charged situations (Zhang et al., 2023). Also, the appearance of robot issuing the recommendation (anthropomorphic vs. mechanical) might influence the acceptance and trust in the provided recommendations (Laakasuo, 2023; Malle et al., 2016).

Using a between-subjects design online experiment with two groups, we asked participants to take a decision in a moral dilemma situation, deciding over life and death. They are supported by either an anthropomorphised robot (Pepper) or a mechanical robot (Temi), who introduces the decision task via a recorded video. People then need to take their decision first. After that, the robot recommends a pretested unlikely (human) choice and participants get to reevaluate their own decision. Several self-report measures on the participant as well as measures evaluating the robot and the decision task are included.

Through this approach, the study contributes to a deeper understanding of human-robot interaction dynamics in moral decision-making scenarios and offers valuable insights for the future design of AI-supported recommender systems.

A potential way to reduce discomfort in human-robot interaction (HRI) is to enhance control over the robot (De Graaf & Allouch, 2013) and foster a sense of ownership (Delgosha & Hajiheydari, 2021). Research suggests that control is linked to psychological ownership: the more control a person has, the more they perceive the object as part of themselves (Prelinger, 1959). However, pre-existing negative attitudes toward robots may alter these effects, making control and ownership less effective for individuals with strong robot skepticism (Nomura et al., 2008).

Accordingly, this work investigates how control and ownership reduce discomfort in HRI, considering negative attitudes toward robots (RQ1).

The Self-Determination Theory (Deci & Ryan, 2000) highlights three psychological needs—autonomy, competence, and relatedness—that are essential for motivation and well-being. These needs should be considered in technology design to enhance user acceptance (Moradbakhti et al., 2022).

Though distinct, control and self-determination are linked (Deci & Ryan, 2000). Control mechanisms over a robot provide users with a framework to influence outcomes, enabling autonomy and competence. Psychological ownership fosters attachment and integration into one’s identity (Pierce et al., 2001), suggesting it may also foster social relatedness with technology.

Therefore, this work investigates how control and ownership influence the fulfillment of psychological needs in HRI (RQ2).

To investigate these questions, participants (N=64) interacted with the Boston Dynamics robot SPOT in a between-subjects experiment with three groups. Participants either observed the robot autonomously navigating an obstacle course (No control or ownership), controlled the robot to complete the course (Control), or were told they had temporary ownership of the robot, and controlled it to follow the course (Control and Ownership). As measurement, we used the RoSAS Scale, Basic Psychological Needs Scale for Technology, NARS Scale, and manipulation checks for control and ownership.

Social robots are increasingly used in healthcare and education to address personnel shortages. In these fields, relationship formation and affective bonding are crucial, fostering trust, improving communication, and enhancing collaboration between professionals and patients or students (Getson & Nejat, 2022; Belpaeme et al., 2018).

This study examines how communication with a social robot supports socio-affective relationship formation over time and how this, in turn, improves communication. Specifically, we examined turn-taking as a fundamental cue for fluid exchanges between interlocutors. Aligning multimodal turn-taking cues (TTC) poses a significant challenge in human-robot interaction due to the complexities of human conversational dynamics and current technical limitations in real-time processing (Skantze, 2021).

We conducted a 3x2 mixed design with TTC (3 levels: (a) both robot and human used explicit TTC, (b) only the robot used explicit TTC, (c) no designed TTC), and Reference mimicry (2 levels: The robot mimicked or did not mimic participants' linguistic references) as between-subjects factor and repeated interactions as a within-subject factor. In a longitudinal lab-experiment, 35 participants played a ‘spot-the-difference’ game with the NAO-robot in a non-scripted conversation. We developed a collaborative game-based framework (Kruijt et al., 2024), based on Clark and Wilkes-Gibbs (1986), and integrated a pragmatic language model to improve interaction dynamics. The game involved three sessions with six rounds each, requiring participants to verbally describe differences between their scene and the robot's scene. We measured the impact of TTC and Reference mimicry on affective bonding, the feeling of having a real conversation, and common ground before and after each session.

Hypotheses: H1: Explicit TTC increases the feeling of having a real conversation (H1a) and affective bonding H1b). H2: Reference mimicry increases affective bonding (H2a) and common ground (H2b).

This study advances Media Psychology by highlighting the role of turn-taking and reference mimicry in fostering human-robot connection.

Podcasts have established themselves as a relevant medium in everyday life, both worldwide and in German-speaking countries. Previous research has identified a variety of motives for podcast listening (e.g., by applying the uses and gratifications approach). However, past research on motives tends to focus on podcasts as a medium. which is far too broad and ignores subcategories, or on specific genres such as true crime, which ignores other genres.

The aim of this qualitative project is to examine podcasts broadly and to look at subcategories at the same time, with a focus on German-language podcasts. These subcategories can be conceptualized in different ways. One common method is to focus on genres, which are usually topic-oriented (e.g. true crime, news). Another method is to focus on the design and production style. During the project, we developed a categorization in which we differentiate between four typical types of design/production styles (“story”, “show”, “factual”, “talk”). These categories were established through listening analyses of podcasts and in discussions with experts.

The proposed talk is dedicated to the question: what are the motives for listening to German-language podcasts, and how are these motives related to different subcategories (topic-oriented genres as well as design/production styles)?

Semi-structured interviews were conducted with 30 regular users of German-language podcasts. Questions focused mainly on one specific podcast per interview and on why interviewees listen to the selected podcast (motives). Podcasts were selected to explore a diverse range of subcategories within the podcast medium (16 different podcasts in total). To investigate the motives within the different subcategories of podcasts, a qualitative content analysis according to Mayring (2010) was carried out using MAXQDA.

We aim to extend the qualitative findings with a quantitative study in the future.

Human-Computer Interaction necessitates a nuanced exploration of Usability, User Experience, and Usage Perception, particularly in contexts characterized by varying levels of perceived technological complexity. This study investigates this statement by examining the distinction between these constructs, the sensitivity of their interrelation, and how interaction, experience and perception shape users’ technology use. A comprehensive online survey was conducted with 313 participants in Germany, who assessed interactions with three distinct technologies ranging from simple devices (e.g., vacuum cleaners) to advanced systems (e.g., artificial intelligence).

The results confirm the discriminant validity of these constructs, demonstrating that Usability, User Experience, and Usage Perception are empirically separable dimensions of Human-Computer Interaction, through and across perceived technological complexity. Importantly, the distinction becomes increasingly pronounced with higher perceived technological complexity.

This highlights the need for Human-Computer Interaction research and technology design to account for the dynamic interplay between interaction, experience, and perception, particularly in complex technological environments.

Advancements in machine learning, particularly Generative Adversarial Networks (GANs), have enabled the creation of highly realistic synthetic images and videos, commonly known as deepfakes. While deepfake technology has applications across various domains, it has been extensively used to generate sexually explicit content of individuals without their knowledge or consent. These non-consensual materials, referred to as deepfake pornography, constitute approximately 98% of all deepfake content, which is reflected in a growing online economy. This raises concerns about victims of deepfake pornography often remaining unaware of their inclusion in such content and lack of control over its distribution, leading to significant psychological distress.

Despite the increasing prevalence and consumption of deepfake pornography, there is little empirical evidence on how victims are perceived by the public. This study addresses this gap by examining the perceived attributed agency and responsibility attributed to victims of deepfake pornography. Using a 2 × 2 vignette study design, we explore the impact of the victim’s gender (female vs. male) and public status (public vs. private figure) on these perceptions. Additionally, we consider the role of participant´s gender, AI literacy, and compassion for the victims as influencing factors.

As deepfake technology continues to advance, enabling the low-threshold creation of sexually explicit content featuring any individual, irrespective of their public status, understanding public attitudes on the agency and responsibility of the victims becomes increasingly critical. This study aims to provide first insights into the public´s perception that can inform media discourse, legal frameworks, and privacy regulations that mitigate the harm caused by non-consensual deepfake pornography.

Social robots offer the potential to enhance primary education, especially in schools facing teacher shortages. While robots have been used to teach typical subjects like mathematics and second languages (Belpaeme et al., 2021), children with special needs, such as those with Autism Spectrum Disorder (ASD), may particularly benefit from robot tutors (Scassellati et al., 2018). Children with ASD have socio-communicative deficits, making interactions with robots easier than with humans (Bartl-Pokorny et al., 2021). Furthermore, they have attention deficits, but robots can engage children more effectively (Van Otterdijk et al., 2020), which is crucial for successful learning outcomes. However, little is known about the needs and preferences of both teachers and children with ASD regarding the use of robots in education, or which type of robots would best fulfill these needs.

This study addressed the question: What are the needs and preferences when using social robots as tutors for children with ASD? A qualitative approach will be used, including semi-structured interviews and focus groups, in preparation for later experiments. Interviews will first be conducted with special education teachers to identify desired tasks and behaviors for the robot, which will be used to design interactions for the focus groups with multiple robots. Subsequent focus groups (n-total=25) involve teachers, parents of children with ASD, and children with ASD aged 7-9. In separate groups, their needs, preferences, and attitudes toward robots in education are examined while interacting with six different humanoid robots. Data will be collected through voice recordings, observation notes, and photographs and analyzed using inductive coding (Braun & Clarke, 2006). Findings further the field of media psychology by offering insights into the design and use of social robots as educational tools for children with ASD, helping to align robot capabilities with the specific needs of educators and children.

This study investigates how privacy concerns impact users' risk beliefs and attitudes toward LLMs. Internet Users’ Information Privacy Concerns (IUIPC) model by Malhotra et al. (2004) has three dimensions: awareness of privacy practices, control, and collection. These privacy concerns have been found to impact behavioural intention, trust belief and risk beliefs. To examine these assertation in the LLM context, the first research question asks: What is the relationship between privacy concerns of using LLMs and risk belief and attitudes towards LLM?

Prior research has shown inconsistencies in the findings of the relationship among behavioral intentions and the behavior that is grounded in privacy concern, which is called privacy paradox. It is plausible that cognitive appraisal might play a role. This study considers critical thinking and AI literacy. It is plausible that people with higher levels of critical thinking skills may be better equipped to understand and evaluate the implications of privacy concerns in AI technologies. Critical thinking assesses individuals' ability to analyze information, make reasoned judgments, and approach problem-solving tasks with a critical mindset. This study therefore proposes that critical thinking may mediate the relationship between privacy concerns and attitudes and beliefs towards LLMs. AI literacy is crucial for enabling people to critically evaluate and responsibly utilize large language models. There is limited research on the specific role of AI literacy in moderating the relationship between privacy concerns and risk perceptions and attitude towards LLMs. The second research question asks: To what extent do AI literacy moderate and critical thinking mediate the relationship between privacy concerns and the perceived risk and attitude of large language model users?

Quantitative online survey will be conducted. Measures include privacy concerns (IUIPC), risk beliefs, AI attitude scale, AI literacy scale, Computational thinking scale.

Attention is a necessary, but limited, resource for engagement with media messages. Consequently, accounting for resource allocation during processing is necessary for understanding how individuals learn from and are persuaded by messages. Although it is agreed that attention is a limited resource for information processing (Marois & Ivanoff, 2005), there is somewhat less agreement as to what these resources are, and how they operate. A dominant theoretical perspective in media psychology suggests that our attention draws from a single, unidimensional pool (see Fisher et al., 2018a; Lang, 2017). In cognitive psychology, however, attentional resources are often conceptualized as two pools: cognitive and perceptual (Murphy et al., 2016). Critically, differentiating cognitive and perceptual resources introduces contrasting hypotheses regarding how attention affects information processing (Lavie et al, 2004). Notably, in the cognitive psychology literature, moderate perceptual load is unique in its ability to focus attention. Thus, (H1) cognitive load should have a negative effect on attention, whereas (H2) perceptual load should have an inverted U-shaped relationship. In response to a call for more communication research which adopts this two pool conceptualization (Fisher et al., 2018b), the present study manipulated cognitive and perceptual load in a persuasive health message and measured resource allocation during message processing via secondary task reaction times (STRTs). Specifically, cognitive load was manipulated using language difficulty, and perceptual load was manipulated using figure-to-background contrast. College student participants (N = 142) read a persuasive health message while responding to three visual STRT probes. STRTs provide evidence for attention allocation during message processing to track load over time. Additionally, persuasion was measured at pre- (1-5 days prior to experiment) and post-test. Only by separately manipulating cognitive and perceptual load, and at three levels, are previously overlooked relationships between load, attention, and persuasion observable.

The effects of violent video games on players' social outcomes have been debated for decades, without clear consensus. While most past research and meta-analyses have focused on aggression (Anderson et al., 2010; Calvert et al., 2017; Ferguson et al., 2020; Greitemeyer & Mügge, 2014; Prescott et al., 2018), a promising approach for advancing this field is to study underlying mechanisms, such as desensitization. Desensitization can be understood as a reduction in emotion-related reactivity upon exposure to media violence (Carnagey et al., 2007, p. 490). This project aimed to systematically review the literature on desensitization with a multi-level meta-analysis. We grouped the possible desensitization-related outcomes observed on brain, body, and behavior levels. This approach allowed us to answer the following questions: To what extent is violent video game exposure related lower emotion-related reactivity that could be interpreted as desensitization? What is the direction of this effect, and what variables moderate it? We considered the following moderating factors: study design, age, gender, and compared results based on the outcome type (psychophysiological, neural, behavioural). This meta-analysis significantly contributes to the ongoing debate on violent video game effects on social outcomes by synthesizing the existing evidence across multiple assessment levels and investigating potential key moderating factors.

Method

This meta-analysis plan was pre-registered at PROSPERO (link left out for a blind review).The current work follows the recommended guidelines for systematic reviews (PRISMA 2020 Checklist; Page et al., 2021).

Inclusion criteria

Studies matching the following criteria were included: (1) published in peer-reviewed sources, (2) titles, abstracts, and full texts were available in English or other Latin languages, (3) involved original quantitative empirical studies (cross-sectional, longitudinal, and experimental), (4) tested a relationship between exposure to violent video games and brain, behavior, or psychophysiological outcomes that could be interpreted in terms of desensitization.

Search

Three databases (Scopus, PsycINFO, and Web of Science) were searched with a timeframe from inception to May 8, 2024. No limitation on publication language was used in the search. The search included keywords and free text terms for (synonyms of) "violent video game" AND (synonyms of) 'desensitization' or other outcomes-related keywords AND (synonyms of) methods-related keywords such as “brain” or “body” or “behavior”.

Screening procedure

Following the search, duplicates were excluded in Rayyan. Next, study selection was performed in Rayyan by three independent reviewers. Any discrepancies between the reviewers were resolved through discussion.

Data extraction

A standardized data extraction form was developed to extract the following data from each article: (1) general information (title, author, publication year, DOI) (2) sample characteristics (country, age, age group, sample size, gender distribution, education level, gaming experience), (3) study design, (4) predictor measurement, including details of manipulation (for experimental designs), (5) outcome measurement (behavioral, neural, psychophysiological).

When available, the following information about the outcome variables was extracted: unit of measurement, M, SD, effect size, and p-value. Three independent reviewers extracted the data, resolving disagreements through discussions with at least one additional reviewer. All data were recorded in an Excel spreadsheet and later cross-checked by at least one more independent reviewer.

Risk of bias (quality) assessment

The methodological quality of the included studies was assessed with 12 items based on the Cochrane Collaboration’s tool (Higgins & Altman, 2008) following the application by Pereira et al. (2023). Following Taneja et al. (2021), reports were later classified as low, moderate, or high quality. The risk of bias analysis indicated that all of the studies were of high quality.

Preliminary Results

The search across the three databases resulted in 936 reports, which were reduced to 542 after deduplication. Screening of full texts identified 80 reports, of which 67 reports were included in the final analysis. So far, the reports on behavioural (n = 12) and physiological outcomes (n = 15) have been analysed, with a total of N = 5,970 participants and 62 effect sizes. Brain outcomes are preparation for the conference presentation.

The summary effect estimates and 95% CIs of included meta-analyses were analyzed using a random effects model. For the behavioural outcomes, the results showed non-significant effect for habitual desensitization, r=–0.07, 95% CI [–0.20, 0.06], SE=.06, t(10)=–1.19, p=.263, non-significant for effect short-term desensitization, r=–0.15, 95% CI [–0.30, 0.01], SE=.07, t(17)=–2.06, p=.055. For physiological outcomes, the results indicated non-significant effect for short-term desensitization, r=–0.08, 95% CI [–0.20, 0.04], SE=.06, t(23) = –1.35, p = .189. The total level of heterogeneity for these effects was high, ranging between I2=21.41% and I2=55.99% for behavioral studies, and I2=0.00% and I2 = 88.34% for the physiological outcomes. The sensitivity analysis indicated no evidence of publication bias.

Discussion

Taken together, these preliminary results based on the behavioral and physiological outcomes do not support the presence of the desensitization effect related to violent video game exposure. Overall, the present findings provided a more nuanced understanding of the violent video game effects on desensitization-related outcomes. Assessing physiological, behaviour (and brain) outcomes separately highlighted the high heterogeneity across studies potentially due to varying measurement approaches. Based on the quality assessment, practical implications for future studies will be discussed. Moreover, theoretical implications will be explored for existing models (Carnagey et al., 2007) and potential new frameworks.

Advancements in technology have revolutionized opportunities for online self-presentation. Editing apps now allow users to enhance and ‘beautify’ their appearance with a simple click, and these idealized images can be shared on visually-driven social media platforms. Previous research has explored how active online self-presentation behaviours influence body image perceptions, with facial satisfaction being particularly prominent (e.g., Tiggemann et al., 2020), as well as cosmetic surgery acceptance and intentions (e.g., Hermans et al., 2022). However, much of this research has treated media effects in isolation, often overlooking the social context in which beauty standards and appearance practices are negotiated. As highlighted by scholars in this field (e.g., Piazzesi & Lavoie Mongrain, 2020), it's crucial to consider the role of peer influences and social norms within this context (Sociocultural Theory; Tiggemann, 2012). Additionally, more nuanced investigations into the mechanisms that link social media use with body image outcomes are needed (e.g., Veldhuis, 2020).

In response, this study explores how active online self-presentation and social norms influence women's facial satisfaction, as well as their acceptance of and intention to undergo cosmetic procedures. It further investigates whether these relationships are explained by variations in users’ self-objectification, and whether they particularly occur in individuals who engage more frequently in upward appearance comparisons with other women online.

To examine the proposed relationships, we conducted a cross-sectional survey of young adult women (N=138; Mage=22.45, SDage=1.58). The key variables included active self-presentation (e.g., selfie editing and posting), perceived social norms regarding cosmetic procedures, facial satisfaction, and acceptance and intention to undergo cosmetic procedures. Additionally, we tested the moderating effect of upward social comparison with other women’s appearances encountered online, as well as the mediating role of self-objectification. The results and their implications for both media effects research and media literacy interventions will be discussed at the conference.

Social media has revolutionized the way we connect, communicate, and consume information. However, its darker aspects pose significant challenges to individuals and societies. This Special Issue (SI) seeks to explore the negative psychological consequences of social media, such as their impact on interpersonal communication, the formation of political beliefs, mental health and body image, consumer behavior, and information processing.