Xiaojiang Zheng

Computers & Education: Artificial Intelligence (CAEAI) 11 p. 100625-100625 Doi: https://doi.org/10.1016/j.caeai.2026.100625

Drawing on expectation confirmation theory (ECT), this study explores the factors that impact Human-Artificial Intelligence (AI) collaboration and investigates the effect of students' distrust towards GenAI. An online survey was administered to students in higher education using GenAI (N = 245). This study found positive and significant relationships between effort expectation and performance expectation, as predicted by ECT, with positive confirmation of GenAI collaboration significantly influencing intentions and positively impacting student behavior. This study further revealed that GenAI distrust negatively moderates the relationships between (1) effort expectation and performance expectation; and (2) expectation confirmation toward collaborating with GenAI and intentions to collaborate with GenAI. Accordingly, this research contributes to the understanding of the role of distrust into Human-AI collaboration and extends the theoretical boundaries of ECT in higher education context. Practically, our findings provide insights for educators and GenAI practitioners to develop strategic approaches that effectively bridge students' distrust and initial expectations with responsible Human-AI collaboration behaviors.

Frontiers in Virtual Reality 7 Doi: https://doi.org/10.3389/frvir.2026.1860842

Research in virtual reality and related immersive environments increasingly relies on avatars as perceptual and embodied carriers of social identity, yet existing avatar systems offer limited control over identity cues, restricted demographic diversity, and weak reproducibility. These constraints narrow the questions that can be asked about embodiment, social categorization, and intergroup processes. We introduce synthetic embodiment as a theory and method framework for studying AI-generated identity cues in virtual reality and related immersive environments, including persistent metaverse contexts. The framework reconceptualizes identity not as a fixed categorical label, but as a multidimensional and controllable design space varying across phenotype, gender presentation, age, cultural markers, ambiguity, and intersectional cue structure. We further propose the Embodiment Categorization Bias (ECB) model, which specifies a testable process through which identity cues operate as perceptual inputs, shape embodiment strength, guide social categorization, and influence intergroup outcomes such as bias, empathy, perceived threat, cooperation, and social distance. To support empirical work, we outline a methodological pipeline for avatar generation, standardization, immersive integration, perceptual validation, and fairness auditing. An illustrative feasibility demonstration shows how theoretically specified identity variables can be translated into structured generative prompts and preliminary avatar stimuli across multiple generative AI systems. This demonstration is not presented as perceptual validation or as empirical evidence for the ECB model, but as an illustration of the early stimulus-construction stage of the synthetic embodiment pipeline. Together, synthetic embodiment positions AI-generated avatars as research infrastructure for more precise, scalable, and ethically accountable studies of identity in immersive environments.

European Conference on Information Systems

In metaverse platforms, avatars are used to represent users' virtual identities. Hence, the aesthetic features of avatar products are a core driver of consumption decisions, yet there is still limited research on how different image aesthetic characteristics influence product sales. Rooted in information processing theory, this research will examine four image aesthetic features, including colorfulness, color complexity, visual atypicality, and visual fluency, and employs fuzzy-set qualitative comparative analysis (fsQCA) to identify how different configurations of these features influence avatar product sales. This study aims to provide novel insights to current information systems literature by revealing unique information processing patterns through which combinations of aesthetic stimuli jointly shape avatar product purchases in metaverse platforms. Practically, this study provides actionable insights for avatar product creators to optimize image designs and enhance the avatar product sales.

International Conference on Management and Organization

Avatar-mediated communication presents a novel avenue for leaders to express their authentic selves through digital embodiment, extending the concept of embodied authentic leadership (EAL). However, the specific mechanisms by which avatars enable users to engage in EAL remain underexplored. To elucidate the underlying psychological pathways from avatars to EAL, this study integrated social identity theory (SIT), which explains how people define themselves based on their membership in various groups (Tajfel, 1978), with a biographical case study of Mats Steen. Through an analysis of Mats's biographical documentary and blog narratives, we identified two key enabling mechanisms: from personal level, avatars facilitate the convergence of Mats's constrained physical self-identity and his idealized virtual self-identity. This convergence enhances his self-efficacy and reduces disinhibition, thereby fostering EAL. From social level, we find that the avatar-enabled depersonalization (shifting from individual to social identity focus) paradoxically catalyses social identification with the group, leading to higher levels of EAL. These findings contribute to leadership theory by proposing a framework for digital embodiment in EAL and challenge traditional, physically cantered models. They also offer implications for management research and organizations to leverage avatars and enhance leadership effectiveness.

International Journal of Contemporary Hospitality Management p. 1-20 Doi: https://doi.org/10.1108/ijchm-05-2025-0741

Purpose This study aims to examine the relationship between place solidarity, hospitality attitudes and mental health outcomes in the context of the Ukraine−Russia conflict. Despite the conflict’s profound effects on global hospitality operations, this is an initial exploration of the influence of place solidarity on hospitality brand attitudes and guest well-being. Design/methodology/approach Structural equation modeling is used to analyze the relationships among place solidarity, hospitality attitudes and mental health outcomes, based on two cross-sectional surveys of 758 US individuals. Findings The two studies show that place solidarity is a double-edged sword by improving hotel brand responses (Study 1) but challenging guests’ well-being (Study 2). Study 1 shows that place solidarity increases perceived brand social responsibility, brand trust and personal reputational benefits while also heightening decisional fatigue. Study 2 identifies the impacts of higher place solidarity on emotional exhaustion, poorer sleep quality, fear of expressing Russia opposition on social media. Practical implications The finding that place solidarity produces dual outcomes in hospitality contexts underscores the need for sensitivity to guests’ psychological states, especially in situations where solidarity and solidarity signaling is happening. Originality/value This research contributes to the hospitality literature by exploring the complex relationships among place solidarity, hospitality attitudes and mental health outcomes during international conflicts. To the best of the authors’ knowledge, this is the first study to show that place solidarity influences both hospitality decisions and psychological well-being in a geopolitical crisis.