This study examines LLM-based AI chatbots as social interaction partners. It identifies young adults’ support-seeking motivations and models their associations with two interaction processes (Perceived Authenticity and Interaction Intensity) and two psychological outcomes (Resilience and AI Dependence). In Study 1, we developed the Motivation Scale for Generative AI (MS-GAI) through theory-driven item generation and interviews and validated it with a survey of Korean young adults (N = 300). Exploratory and confirmatory factor analyses supported an eight-factor, 40-item structure grounded in four social-support domains and supplemented by context-specific motivations. In Study 2, structural equation modeling showed that Emotional Support and Self-Enhancement were positively associated with both Perceived Authenticity and Interaction Intensity. Objective Advice Seeking was positively associated with Perceived Authenticity, whereas Learning Assistance, Daily Problem-Solving, and Entertainment/Mood Diversion were positively associated with Interaction Intensity. Perceived Authenticity was also positively associated with Interaction Intensity. Both interaction processes were positively related to self-reported Resilience and AI Dependence, with Interaction Intensity showing the stronger association with AI Dependence. These findings identify distinct but overlapping interaction pathways through which support-seeking motivations are associated with adaptive and maladaptive outcomes. The study extends understanding of social support in human–AI interaction by providing a validated motivation measure and showing how shared interaction processes relate to both short-term perceived recovery and dependence. The findings are consistent with the Paradoxes of Technology perspective and highlight the need for emotionally supportive yet balanced design and policy safeguards.
Blockchain technology is increasingly integral to domains such as finance, supply chain, commerce and tourism, yet consumer trust remains a pivotal barrier to its widespread adoption. Studies on trust in blockchain contexts have significant conceptual and contextual limitations that inhibit a holistic understanding of trust formation. In this study, we draw on the antecedents, decisions, outcomes and theory, context, methods (ADO-TCM) framework to systematically review 53 high-quality multidisciplinary studies. We identify three critical issues: (1) technological determinism that conflates technical capability with consumer trust, marginalising psychological, social and institutional factors; (2) theoretical convergence around acceptance models that obscures blockchain-specific trust phenomena; and (3) a narrow geographic and domain scope that creates self-reinforcing knowledge hierarchies limiting generalisability across applications and cultures. To address these gaps, we propose a multidimensional trust framework integrating three dimensions: technological architecture trust, information source trust and social actor trust, reflecting blockchain’s sociotechnical structure. Central to this framework is a four-paradigm typology that classifies trust into interpersonal, technology, institutional, and sociotechnical perspectives, revealing that the prevalence of technological determinism reflects theoretical inertia rather than genuine consensus. We complement this with three research themes: multi-level theoretical integration, innovative methodologies and contextual expansion. This study contributes to the information systems discourse by providing a theoretical foundation for understanding trust in complex sociotechnical systems beyond acceptance models. By conceptualising consumer trust as a dynamic, multidimensional construct shaped by interactions between users, technology, and institutions, we offer a research agenda with actionable directions for advancing blockchain trust theory and practice.
Emotion-aware conversational systems are increasingly adopted in healthcare, yet most emotion detection approaches rely on sentiment polarity, which fails to capture emotional nuance and intensity in short, informal chat messages. To overcome these limitations, this study develops an interpretable emotion-aware framework called Word Intensity Scoring Engine (WISE) that explicitly models emotional intensity in short-form conversational text. Grounded in Plutchik’s Wheel of Emotions, WISE integrates lexicon-based emotion labeling, synonym expansion, semantic similarity using word embeddings, and negation-aware feature extraction to calculate intensity-weighted emotion scores at the word level. The WISE framework is evaluated through experimental benchmarking and an expert review involving practicing healthcare professionals. Results show that WISE enhances the performance of ML and LLM-based classifiers. Further, expert evaluations indicate that WISE-derived responses are perceived as professional, contextually appropriate, and empathetic. This study contributes a transparent, theoretically grounded design artifact that advances affective computing research and informs the deployment of emotion-aware conversational agents in healthcare.
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.
