Artificial intelligence (AI) is becoming increasingly embedded in employment interviews. Although AI interviewers may enhance the efficiency and standardization of personnel selection, their effects on applicants’ authentic and inauthentic self-presentation remain insufficiently understood. Drawing on impression management theory, this study examines how AI interviewers, relative to human interviewers, shape applicants’ self-presentation and why differences arise. Across four studies, we find that applicants facing AI interviewers perceive fewer opportunities to perform than those facing human interviewers. This reduced opportunity to perform, in turn, weakens not only exaggerated self-enhancement but also authentic self-presentation and authentic self-enhancement. These findings indicate that AI interviewers do not constrain deceptive impression management. They also dampen the expression of genuine strengths and self-relevant information. The results also show that contingency-based interactivity can mitigate this disadvantage by enhancing applicants’ perceived opportunity to perform and narrowing the gap between AI and human interviewers. By identifying both the constraining effects of AI interviewers and the design conditions that can reduce them, this study advances the research on AI interviews and offers practical implications for the design of more effective AI-enabled interview systems.
Artificial intelligence (AI) plays an increasingly important role in marketing, acting as both a product designer and an endorser, yet research predominantly examines these roles in isolation. This study addresses this gap by investigating how consumers’ adoption of AI-designed products is jointly shaped by textual designer disclosure (i.e., AI- vs. human-designed) and visual endorser cues (i.e., human-like vs. animated avatars). This research proposes that consumers integrate these two modalities into a single evaluation, and their adoption intention is shaped by the congruence between textual and visual cues. Drawing on schema-congruity and curiosity theories, this research tests a designer-endorser congruence effect across four studies. Study 1 (N = 240) demonstrates the congruence effect: for an AI-designed product, consumers show higher adoption intention when it is endorsed by an animated (vs. human-like) endorser, whereas the effect is reversed for human-designed products. Study 2 (N = 156) illustrates that perceptual curiosity serves as a mechanism: AI’s involvement in design stimulates curiosity, which is then sustained by presenting the endorser in a congruent, animated form. Study 3 (N = 156) validates the mediating effect of perceptual curiosity alongside cognitive mechanisms, i.e., perceived trustworthiness and uncanniness. Study 4 (N = 340) examines how this congruence effect varies across consumers with different dispositional traits and shows that this effect is stronger among those with high anxiety about AI. These findings make a theoretical contribution by introducing designer-endorser congruence as a determinant of product adoption in AI-mediated marketing, and by identifying curiosity as an alternative affective mechanism that complements prior information-processing perspectives. In practice, this study provides insights for leveraging avatar endorsement to induce curiosity and to encourage the adoption of AI-designed products.
The top management team (TMT) constitutes a critical driver of successful digital transformation in manufacturing enterprises. Although studies have investigated the relationship between the TMT and digital transformation, research on the influence of dynamic interactions among members—particularly team conflicts—on this relationship remains limited. Drawing on upper echelons theory and organizational learning theory, this study examines the relationship between TMT conflict and digital transformation performance, thereby extending the research on the impact of the TMT on digital transformation. An empirical analysis of data obtained from 358 firms reveals that TMT task conflict enhances digital transformation performance, whereas TMT relationship conflict has a detrimental effect in this context. Furthermore, ambidextrous learning serves as a mediator underlying the influence of TMT conflicts on digital transformation performance. Additionally, digital orientation positively moderates the relationship between ambidextrous learning and digital transformation performance, thereby strengthening the beneficial impact of ambidextrous learning. We discuss the theoretical and practical implications of these findings.
Sleep tracking has become a prominent application area of self-tracking. However, sleep tracking often fails to lead to the desired improvement in users’ sleep habits. Hence, user compliance with the advice from sleep-tracking technology is important for achieving the potential health benefits. This study investigates the factors and causal patterns associated with high and low levels of advice-compliance behavior among users of wearable technology for sleep tracking. We employed a mixed-methods approach consisting of an interview study and a survey. In Study 1, using thematic analysis, we identified four technology affordances of sleep tracking: quantifying sleep-related data, obtaining sleep-related guidance, triggering behavioral changes, and socializing. In addition, we identified three psychological outcomes associated with affordance actualization through user responses: pursuing perfect results, raising awareness of sleep health, and limiting attention to sleep tracking. In Study 2, we used fuzzy-set qualitative comparative analysis to examine how different configurations of technology affordances and psychological outcomes contribute to advice-compliance behavior. The results indicate that obtaining sleep-related guidance, triggering behavioral changes, and pursuing perfect results are important elements in configurations associated with a high level of advice-compliance behavior. This study contributes to research and practice by providing new empirical insights into technology affordances and associated psychological and behavioral outcomes, offering suggestions for technology design.
AI is increasingly used to improve Web3 infrastructure, smart contracts, and decentralized applications, but its effects are not only technical. This paper develops a socio-technical framework for examining AI in Web3, grounded in the Task, Technology, People, and Structure dimensions of Leavitt’s Diamond. Through a systematic review of 370 studies, we analyze how AI reshapes Web3 across the infrastructure, execution and logic, and application layers. We show that AI expands operational tasks, introduces new technical dependencies, redistributes expertise and decision-making, and alters governance structures such as incentives, decentralization, privacy, and accountability. Across the three layers, we identify six research agendas: the task complexity spiral, the decentralization-capability trade-off, role reconfiguration and capability inequality, the diagnostic-governance response gap, the accountability gap, and privacy erosion through asymmetric monitoring. The paper contributes a cross-layer socio-technical framework for evaluating AI for Web3 and outlines directions for building systems that are not only technically effective but also institutionally accountable.
Digital platforms face a fundamental paradox: while expanding service variety is a dominant competitive strategy, it risks inducing a “paradox of choice” that confuses and deters users. This tension manifests with extreme clarity in the nascent, high-complexity market of cryptocurrency exchanges, creating a pressing empirical puzzle. To resolve this, we adopt the Stimulus-Organism-Response (SOR) perspective in a three-stage mixed-method study to investigate how platforms can strategically manage this trade-off. Our qualitative exploration (Study 1) established a capital flow schema called “inflow, roll, and go” and identified key complexity-reduction mechanisms. A subsequent survey (n = 190, Study 2) validated that perceived innovativeness and scalability are critical stimuli for service variety, which in turn drives user continuance intention. A final survey (n = 140, Study 3) confirmed that users prioritise services that bridge to the traditional financial system, forming a minimal viable structure with a variety of functions. Our meta-inferences make several key contributions, including the resolution of the service variety paradox by introducing a theoretical distinction between value-adding “real-variety” and confusing “pseudo-variety” and the development of a strategic roadmap that guides exchanges in navigating the tension between service expansion and user confusion, offering actionable insights for platform strategy in any high-velocity digital market.
Grounded in the resource-based theory, this research investigates how tangible, human, and intangible digital technology (DT) resources shape the digital business capability of organizations. To capture both linear and configurational effects, a multi-method approach integrating partial least squares structural equation modeling (PLS-SEM) with fuzzy-set qualitative comparative analysis (fsQCA) was employed. The outcomes of PLS-SEM indicate that digital technology infrastructure and digital employee skills significantly enhance digital business capability, whereas digital organizational culture alone does not exert a direct effect. DT–business alignment weakens the positive influence of digital technology infrastructure, suggesting that overly rigid alignment may constrain the benefits of digital investments. Complementing these findings, the fsQCA results identify multiple equifinal resource configurations through which firms can develop digital business capability under different levels of DT–business alignment and government digital support. Overall, this study advances theoretical understanding of how digital resources collectively contribute to capability formation and offers practical guidance for firms seeking to allocate resources strategically and strengthen their competitive advantage.
Generative artificial intelligence enables the rapid production of fluent, plausible, and readily shareable content, but it also increases the likelihood that problematic information will circulate before it can be adequately verified. In such environments, correction is consequential not only because it may challenge questionable content, but also because it may signal how responsibility for problematic information is being governed. Building on this premise, this research examines how visible correction shapes user responses to generative AI hallucinations by comparing platform-issued correction and community-based correction. We argue that these two corrective actors are analytically important not because they are mutually exclusive in practice, but because, within a focal visible correction episode, they are more likely to direct users’ responsibility judgments toward different actors. Across three experiments in an investment-information context, we examine whether, why, and when visible corrective actors matter. Study 1 establishes baseline differences among no correction, platform-issued correction, and community-based correction. Study 2 examines governance responsibility interpretations while accounting for a negative-affect alternative account. Study 3 tests correction strength as a boundary condition. The findings show that visible correction reduces willingness to further disseminate problematic content, whereas platform-issued correction more consistently strengthens trust in platform governance. By reconceptualizing visible correction as a governance-relevant signal, this research extends correction research in generative AI contexts and distinguishes between platform-focused governance evaluation and downstream diffusion response.
Existing research has not examined how embedding algorithms into public sector performance evaluation systems influences street-level bureaucrats’perceptions of algorithmic authority and compliance tendency. Drawing on technology enactment theory, this study uses an experimental design to examine the effects of algorithm embedding degree and algorithm transparency on algorithm compliance in public performance assessment, while semi-structured interviews provide triangulation, mechanism illustration, and contextual explanation for the experimental findings. The results show that higher embedding and greater transparency significantly strengthen bureaucrats’algorithmic compliance tendency. Under high embedding and low transparency, algorithms are viewed as the core evaluative mechanism, yet unclear rules lead bureaucrats to rely on superior judgment, producing compliance with both algorithmic and leadership authority. Under low embedding and low transparency, weak algorithm dominance causes compliance to revert to leadership authority. Perceived predictability of evaluation outcomes partially mediates the relationship between algorithm embedding and compliance tendency. Overall, the study reveals how algorithm embedding influences authority-related compliance orientations and provides empirical evidence for understanding authority-related changes in public performance appraisal.
Privacy cynicism is a pervasive and complex phenomenon that often leads users to ignore privacy concerns and inhibits privacy-protective behavior. However, few studies have proposed or empirically tested design mechanisms aimed at mitigating privacy cynicism. To address this gap, we integrate design science with experimental and survey research to examine the influence of the design of dual lists, an emerging privacy‑notice mechanism mandated in China on privacy cynicism reduction during app use. Drawing on Construal Level Theory (CLT), we identify and implement three dual-list design features, namely data sharing occurrence, data type label, and dual-list salience, which are theoretically linked to three distinct dimensions of psychological distance. Using the Elaboration Likelihood Model (ELM), we further propose distinct routes through which these features influence users' privacy cynicism. Results show that data sharing occurrence and data type label mitigate users' privacy cynicism by reducing privacy uncertainty via the central route of information processing, both independently and through their interaction. Dual-list salience mitigates privacy cynicism through the peripheral route, ultimately shaping users' privacy-protective behavior. Interestingly, dual-list salience is effective in mitigating privacy cynicism only when the occurrence of data sharing is disclosed. We also find that privacy uncertainty exerts a double-edged effect on users' privacy-protective behavior through a competitive mediation mechanism. Theoretically, this study advances privacy cynicism research by shifting the focus from explanation and prediction to theory‑driven, design-based interventions. Practically, this study offers actionable insights for users, service providers, and policymakers seeking to foster a more trustworthy and healthier digital ecosystem.
The literature on artificial intelligence (AI) crafting has focused predominantly on its effects on individuals themselves. However, a crucial yet overlooked issue persists in understanding how coworkers' AI crafting affects observers. Drawing on affective events theory, this study develops a model to examine when and how coworkers' AI crafting influences observers' emotions and behaviors. Using a 2 & times; 2 between-subjects situational experiment (Study 1) and a multi-wave field survey (Study 2), the findings reveal that when observers and coworkers share high cooperative goal interdependence, observers are more likely to feel admiration for coworkers' AI crafting, and in turn to engage in observational learning behaviors. Conversely, when observers and coworkers have high competitive goal interdependence, observers tend to experience contempt for coworkers' AI crafting and are more inclined to engage in workplace ostracism. These findings broaden our understanding of the impact of AI crafting from the perspective of third-party observers and provide practical insights for organizations seeking to manage and leverage AI crafting effectively.
While service robots are transforming industries, existing studies often assume a high degree of immunity to the influence of situated physical environments. This assumption may not hold in the context of robotic services, in which physical environments (where a service robot operates, hereafter, physical service context) contribute significantly to service design. This study employs the Privacy Calculus Theory to develop a theoretical framework for interpreting the impacts of physical service context on robotic service adoption. It investigates the influence of anthropomorphism on usage intention through the dual pathway of perceived privacy invasion and trust across three distinct contexts: home, workspace, and public spaces. By a large-scale scenario-based experiment involving 3893 participants, we identified three non-linear relationships: a fluctuating trajectory between anthropomorphism and perceived privacy invasion, a non-linear positive monotonic relationship between anthropomorphism and trust, and an uncanny valley (UV)-like pattern between anthropomorphism and usage intention. The results reveal that usage intention varies significantly across physical service contexts, being lowest at home. The influences of both trust and perceived privacy invasion on usage intention are highly context-dependent. These findings reconcile conflicting results in the literature by demonstrating that the effect of anthropomorphism is non-linear and context-bound. The study contributes by highlighting the critical role of physical service context and privacy perceptions for embodied artificial intelligence (AI) and offers clear guidance for designing and deploying service robots that balance human-like features with privacy-sensitive adoption strategies.
Cross-border higher education alliances implementing micro-credentials face significant challenges in coordinating data governance across diverse institutional and regulatory contexts. While prior research has examined governance frameworks and compliance mechanisms, limited attention has been given to how data governance arrangements evolve in multi-organizational settings over time. This study adopts a process-oriented perspective to examine how data governance evolves within a European higher education alliance and how institutional pressures shape shifts from decentralized to centralized governance. Drawing on institutional theory and a qualitative case study, we analyze how regulative, normative, and cultural-cognitive pressures erode the viability of decentralized governance arrangements, prompting their reconfiguration. The findings show that decentralized governance, while initially aligned with institutional autonomy, becomes increasingly difficult to sustain under competing pressures for standardization, interoperability, and regulatory compliance. In response, a move toward centralized governance unfolds, which enhances legitimacy and coordination but introduces new tensions related to local practices and regulatory diversity. These tensions, in turn, generate ongoing adjustments, highlighting governance as a dynamic and continuously evolving process rather than a static design choice. This study contributes to institutional theory by demonstrating how governance arrangements evolve in inter-organizational contexts under overlapping pressures. It further contributes to data governance research by advancing a process-based understanding of how alliances reconfigure governance in response to shifting institutional demands.
In the framework of Persuasive Systems Design, diet-related apps often connect users with peers or healthcare experts as the indirect and direct routes of persuasion to retain users and promote healthy behaviors. This connectivity reaches broad and diverse populations, yet its causal effects on users’ eating behaviors in digital environments, particularly those centered on photo-based interactions, remain unclear. Drawing on social norms theory, media richness theory, and social comparison theory, we examine the effects of two socially delivered features—photo-informed expert feedback and photo-based peer observation—through a four-month randomized field experiment with 337 participants, orthogonally manipulating their presence in a photo-based healthy eating app. We find that photo-informed expert feedback increases app usage and slows disengagement over time, consistent with the influence of injunctive norms and the richer information conveyed through meal photos. In contrast, photo-based peer observation reduces engagement and has no significant aggregate effect on food choices, consistent with predictions based on the similarity hypothesis of social comparison and defensive avoidance in photo-based environments. We further find that this disengagement mainly occurred among relatively unhealthy eaters. This study advances theoretical understanding of how socially delivered direct and indirect routes of persuasion operate in photo-based, computer-mediated environments. It further provides causal, feature-level evidence from a longitudinal randomized field experiment and offers practical guidance for designing mHealth apps that improve user retention and promote healthy eating behaviors.
This study examines the firms’ combined implementation of Robotic Process Automation (RPA) and Artificial Intelligence (AI) and its association with firm performance using two survey datasets covering large firms worldwide. RPA automates repetitive tasks whereas AI supports cognitive tasks. While previous research typically examines these technologies separately, we draw on complementarity theory to hypothesize that their joint implementation is associated with performance outcomes that exceed the independent benefits of each technology. We find that joint implementation is positively associated with additional revenue growth (linked primarily to price increases) but is not associated with additional cost reduction beyond the separate effects of RPA and AI. These findings challenge the prevailing view that combining RPA and AI technologies primarily enhances performance through increased efficiency, suggesting instead that joint implementation is associated with revenue enhancement. We also consider firms’ strategic intent for adopting these technologies and find that among firms implementing both RPA and AI, a stronger revenue growth intent is associated with greater revenue-related outcomes. Our findings provide a baseline analysis of automation complementarities in the pre-Generative-AI era, even as advancements in technology continue and raise new questions.
This study examines how network position relates to institutionalization and risk perception during organizational adoption of large language models (LLMs). Drawing on Actor-Network Theory and neo-institutional theory, we conceptualize adoption as a process of network stabilization and operationalize translation practices for quantitative analysis. Survey data from 981 employees in Chinese public-sector organizations show that translation practices are positively associated with perceived institutionalization and negatively associated with perceived risk emergence. Perceived institutionalization, however, amplifies rather than buffers this relationship. When perceived institutionalization is low, translation practices show no significant association with lower risk; when it is high, the association becomes substantial. The findings point to an institutionalization paradox: arrangements that stabilize LLM use may also sharpen risk differentiation between central and peripheral actors.
Atmospheric cues play significant roles in e-commerce live streaming (ELS). With "Clear Mode" removing interface elements in ELS, the role of non-interface atmospheric cues has become critical, yet their impacts remain underexplored. Drawing upon the elaboration likelihood model and the dual model of environmental perception, this study develops a research model to examine how contextual environments (CE), a kind of noninterface atmospheric cue, influence consumer attention, perception, and behavior in ELS. Adopting real-world live streaming videos as stimuli, we conduct a 2 (CE: high versus low) & times; 2 (product involvement: high versus low) within-subject eye-tracking experiment to test the research model. The results show that compared to lowCE (LCE), high-CE (HCE) captures more exogenous attention (i.e., attention to the surroundings) and has spillover effects on endogenous attention (i.e., attention to the streamer and the focal product). Exogenous and endogenous attention compete but addictively contribute to perceived diagnosticity, which further increases approach tendency and purchase intention. Additionally, the positive effects of HCE (versus LCE) on endogenous attention are more salient when the product involvement is low (versus high). This study undertakes a pioneering investigation into the role of CE, shedding light on the importance of non-interface atmospheric cues in ELS research. It has practical value for managers, streamers, and platforms in designing live streaming rooms.
People often struggle to regulate their digital technology use, and many existing interventions, such as screentime limits and digital detox programs, rely on external restrictions with mixed and often short-term effects. This research extends research on technology overuse by examining how consumers experience technology disengagement contingent to active mindsets of self-regulation: either as externally imposed and loss-based or as self-endorsed and gain-based. In particular, we examine through the lens of the Joy of Missing Out (JOMO), conceptualized as a positive mindset toward digital disengagement, in contrast to the Fear of Missing Out (FOMO). Drawing on self-determination and regulatory focus theories, we propose that JOMO enhances state authenticity, which in turn strengthens perceived technology freedom and supports digital self-regulation. Across four studies using a multi-method research design (including a text-mining study of 210,706 YouTube comments, two preregistered experiments, and a two-week field intervention), we find convergent support for this account. Findings suggest that JOMO increases state authenticity, which in turn enhances perceived technology freedom, digital well-being, and lowers social media usage. These findings contribute to research on digital well-being, authenticity, and technology overuse by showing that digital disengagement depends on specific selfregulatory mindsets and is more effective when it is experienced as self-endorsed (JOMO) rather than externally imposed (FOMO). Practical interventions for mitigating digital dependency using JOMO (vs. FOMO) mindsets are also discussed.
As generative artificial intelligence (GenAI) reshapes the educational landscape, understanding pre-service teachers’ multifaceted attitudes toward its use is critical. Grounded in the transactional theory of stress and coping, this study aims to identify distinct profiles of pre-service teachers’ cognitive and affective attitudes toward GenAI use, investigate how technostress predicts profile membership, and examine differences in coping strategies, GenAI-enabled productivity and innovation, and teaching commitment across these profiles. Using data from 1895 pre-service teachers in China, we identified three profiles: Cautious Optimists, Ambivalent Adopters, and Highly Aroused Complex Adopters. Higher levels of techno-overload and techno-insecurity were positively associated with membership in the Highly Aroused Complex Adopters profile, whereas techno-complexity and techno-uncertainty were more strongly linked to membership in the Cautious Optimists profile. Regarding outcomes, Highly Aroused Complex Adopters reported the highest levels of both approach- and avoidance-oriented coping, as well as the highest GenAI-enabled productivity, innovation, and teaching commitment. Cautious Optimists exhibited the lowest levels of avoidance-oriented coping alongside moderate GenAI-enabled productivity, innovation, and teaching commitment. Conversely, Ambivalent Adopters reported moderate avoidance-oriented coping but the lowest levels of approach-oriented coping, GenAI-enabled performance, and commitment. These findings underscore the necessity of differentiated teacher education that moves beyond a “one-size-fits-all” approach to training to help pre-service teachers harness the potential of GenAI for sustainable professional growth.
Digital transformation (DT) is a pervasive phenomenon with substantial impact across organizational and societal levels, yet current scholarship largely centers on for-profit contexts where resource ownership is implicitly assumed. This leaves a critical gap in understanding how not-for-profit (NFP) organizations with critical societal impact, which are characterized by resource dependence, heightened accountability, and mission-driven goals, navigate DT. Drawing on a grounded qualitative case study of a large faith-based NFP in New Zealand, this study develops the concept of resource fluidity, an organizational capability to flexibly access, recombine, and redeploy non-owned resources across tasks and time with minimal friction. Resource fluidity unfolds through a three-phase process: (1) Identification of resources with polymorphic potential, (2) Activation through envisioning, redeployment, and experimentation capabilities, and (3) Application guided by mission-driven governance. This study contributes to DT research by shifting the focus from proprietary resource to flexible access-based arrangements and offers a contextualized understanding of how NFP organizations can enact cost-effective and mission-aligned DT under resource constraints.
