Earticle

Rapid advances in mobile technology have accelerated the adoption of personal financial management (PFM) apps and sparked scholarly interest in their influence on consumers’ financial decision‑making. Although prior research has linked financial literacy to prudent financial management behaviors, psychological antecedents—especially motivation—have received comparatively little attention. Drawing on self‑determination theory, this study distinguishes between intrinsic and extrinsic motivation for financial management and examines how each form of motivation shapes PFM behaviors and, ultimately, financial well‑being. Additionally, we explore how specific finance app features—management and external control—moderate the motivation–behavior link. A cross‑sectional survey of 600 active PFM app users reveals that both intrinsic and extrinsic motivations are positively associated with PFM behavior, with app features strengthening the impact of intrinsic motivation toward financial management. This study contributes to the literature by highlighting motivation as a key predictor of PFM behavior and demonstrating the contingent role of specific app features in translating motivation into action. Practically, the findings guide financial service providers in tailoring feature sets to users’ motivational profiles, thereby fostering more effective financial management practices and promoting consumer financial well‑being.

Online review platforms play a critical role in shaping consumer decisions but are vulnerable to review bombing, where coordinated users distort product ratings for non-quality-related reasons. To mitigate this, Metacritic introduced a 36-hour waiting period before users could post reviews for newly released products. This paper investigates the impact of this anti-review-bombing intervention on users’ behavioral responses. Using a panel dataset of over 3,000 products released within nine months before and after the policy intervention, we conduct an empirical analysis and find that the review volume significantly increases after the policy is implemented. More importantly, the average user ratings significantly decreases, suggesting a shift toward more negative sentiment. Our robustness tests, including propensity score matching (PSM) and difference-in-differences (DID), generate qualitatively similar results. These findings highlight the trade-offs between timely participation and user sentiment in platform governance. The study provides empirical evidence on the complex dynamics of anti-review-bombing measures, offering insights for platform designers and policymakers concerned with digital trust and user engagement.

This study examines how generative AI literacy influences users’ intentions to use generative AI through value- based evaluation processes. Conceptualizing AI literacy as a multidimensional capability comprising utilization ability, critical evaluation, ethical use, and creative utilization, the study proposes that literacy affects behavioral intention indirectly by shaping benefit-based (perceived usefulness, hedonic motivation) and cost-based (effort expectancy, performance risk) evaluations, which are subsequently integrated into perceived value. Using survey data from users with prior experience with generative AI and structural equation modeling, the results show that AI literacy exerts differentiated effects rather than uniformly enhancing positive perceptions. Utilization ability and creative utilization increase perceived benefits, whereas critical evaluation reduces perceived benefits and increases perceived risk. Ethical use primarily reduces perceived risk. Perceived usefulness and hedonic motivation strongly increased perceived value, whereas effort expectancy had a weak negative effect. Notably, performance risk positively influenced perceived value, suggesting that among experienced generative AI users, risk awareness may reflect a more informed and manageable evaluation process rather than a purely avoidance- oriented response. Perceived value, in turn, strongly predicted the intention to use. These findings demonstrate the predictive validity of generative AI literacy and highlight its role in shaping user evaluations and adoption decisions.

Virtual try-on (VTO) applications are gaining popularity in fashion industry globally, including jewelry industry. Online retailers are encouraging customers to try these apps for enhanced shopping experience. Despite the growing demand and importance, studies exploring the role of VTO applications in Indian artificial jewelry industry are rare. This study leverages S-O-R model and examines the influence of social and utilitarian cues (stimuli) of VTO on customer attitude (organism), ultimately affecting usage intention of VTO among consumers in the Indian artificial jewelry domain. To get more profound insights, the moderating effect of ‘consumer innovativeness’ was also probed. Data were collected from 368 participants and analysed through Smart PLS 3.0 software. The study findings revealed that perceived usefulness, perceived influence and social proof significantly influence consumer attitude towards VTO, except perceived authenticity. Moreover, customer attitude towards VTO significantly influence their usage intention. Additionally, the moderating effect of consumer innovativeness was found significant, thereby strengthening the relationship between customer attitude and usage intention towards VTO applications. The current research adds value to the online literature by broadening the knowledge on VTO apps usage intention and expanding the application of S-O-R framework in the Indian jewelry fashion industry, an area that has largely remained unexplored.

This study examines the psychological and technology-enabled mechanisms underlying multihoming intention in the over-the-top (OTT) service market, where users increasingly maintain multiple platform subscriptions simultaneously. Grounded in Uses and Gratifications (U&G) theory, this study conceptualizes parasocial interaction (PSI) with actors appearing in OTT content as a core psychological mechanism linking actor-related factors and OTT service characteristics to usage satisfaction and multihoming intention. The model incorporates self-congruity and value congruence as individual-level factors, along with perceived serendipity and telepresence as technological characteristics that enhance parasocial interaction via flow. Survey data from 150 OTT users who had a favorite actor appearing in OTT content and prior experience engaging in fan-related activities were analyzed using partial least squares structural equation modeling. The results indicate that self-congruity and value congruence with favored actors significantly strengthen parasocial interaction. Perceived serendipity and telepresence positively affect flow, which in turn fosters parasocial interaction. Parasocial interaction positively influences both usage satisfaction and multihoming intention, while usage satisfaction further reinforces multihoming intention. This study advances OTT research by shifting the focus from single-platform continuance to multihoming as a strategic usage behavior and by identifying parasocial interaction as a key psychological driver of multihoming intention, offering practical implications for designing technology-enabled flow experiences and relational engagement strategies.

Digital transformation has become a central priority for governments worldwide, aiming to enhance governance and public service delivery. In Indonesia, the Ministry of Home Affairs introduced the Regional Government Information System (SIPD) as a mandatory platform to integrate planning, budgeting, and financial management processes across all local governments. While prior research has predominantly focused on voluntary technology adoption, limited attention has been given to the success of mandatory information systems. This study addresses this gap by examining SIPD using an integrated framework that combines the DeLone and McLean Information Systems Success Model with the Technology–Organization–Environment (TOE) framework, while incorporating user competence as an individual-level capability. Data were collected from 210 local government employees and analyzed using Structural Equation Modeling (SEM). The results indicate that technological factors and user competence significantly influence user satisfaction, which in turn strongly determines system success. In contrast, organizational factors were found to have no significant effect on user satisfaction. Environmental factors, however, have a direct impact on system success. These findings suggest that, in mandatory environments, system success depends more on effective system utilization and user satisfaction than on organizational support alone. This study contributes to the information systems literature by demonstrating how system success mechanisms operate in mandatory-use contexts, where usage is enforced rather than voluntarily adopted, and by integrating TOE and IS Success perspectives to explain both contextual conditions and system outcomes.

Generative AI technologies like ChatGPT have rapidly gained global attention, yet research on their practical applications and user perceptions remains limited. This study addresses these gaps by investigating the factors that influence ChatGPT usage, aiming to enhance the development and adoption of generative AI services. ChatGPT’s characteristics are categorised into information, service, and personal dimensions. Empirical analysis shows that integration and timeliness positively impact perceived usefulness and ease of use, while accuracy does not significantly affect either factor. Convenience enhances perceived ease of use but not usefulness, while compatibility unexpectedly exerts a negative effect on perceived usefulness. Playfulness strongly increases perceived usefulness, whereas self-efficacy unexpectedly reduces it. Differences between experienced and inexperienced users are also identified, offering deeper insights into user behaviour. By clarifying the key drivers and barriers to ChatGPT acceptance, this study provides important academic contributions and practical strategies for developers and organisations seeking to maximise the potential of generative AI technologies.

This study examines how online behavior changed following the April 2022 introduction of IP-based geolocation disclosure on Sina Weibo. Unlike full identity disclosure, geolocation disclosure reveals an identity-relevant but indirect cue, making users partially identifiable while preserving substantial anonymity. Using 2,245,780 comments associated with 36,470 posts, we analyze post-policy changes in comment sentiment, comment length, comment volume, and repost volume. Observed comment sentiment is more positive after implementation, with larger post-policy increases in institutional-account contexts and among verified commenters. We find no statistically significant changes in comment length, comment volume, or repost volume. Horizon-specific estimates indicate that several responses are concentrated near implementation or evolve non-monotonically. Overall, the evidence is clearest for expressive tone. The sentiment response varies across interaction contexts, while several outcomes exhibit temporal heterogeneity.

As organizations explore generative AI for customer-facing and support processes, task-specialized adaptation can improve downstream performance but weaken general conversational capabilities inherited from the base model. This trade-off is especially salient for small language models (SLMs), which are attractive for cost, latency, governance, and deployment flexibility but remain capacity-constrained on reasoning-intensive tasks. This study develops and evaluates a design-oriented stepwise post-training framework for calibrating a 1B-scale language model for bounded business applications while maintaining usable conversational quality. Using Llama-3.2-1B as the student model, the framework integrates warm-up supervised fine-tuning, stepwise reinforcement learning, rationale-consistent distillation, and selective reasoning control. The post-trained model was evaluated on customer intent inference, conversational preference, and selected PISA mathematical-reasoning items. Results suggest improvement over the base model and the same observed Pass@3 accuracy as Llama-3.2-3B on selected PISA mathematics items, with no clear evidence of degradation in limited human-preference tests. The observed gains were task-dependent, appearing more clearly in reasoning-oriented tasks than in routine classification. The study frames compact models as deployment-aware business artifacts and identifies a design logic for balancing task specialization, conversational retention, and resource-conscious model design.

This study examines the potential of the 7C framework as a common organizing structure for review-based UX monitoring in mobile vertical commerce apps. We develop and evaluate the model-development phase of a staged pipeline that identifies UX-related reviews, extracts topics, maps them to 7C categories, and trains a Korean pretrained language model (KcBERT) to reproduce the resulting label structure at the individual-review level. Among 51,505 Google Play reviews from ten apps, manual labeling of 5,100 reviews followed by Naïve Bayes classification yielded 8,260 UX-related reviews (16.0%). LDA produced 11 interpretable topics, including login errors, server speed, delivery/customer support, and product search/filtering. These topics were mapped to 7C categories to generate rule-based training labels. Against these labels, KcBERT achieved 66.16% accuracy and a macro F1-score of 62.81%, outperforming a TF-IDF logistic regression baseline. In a separate human validation of 420 UX-related reviews, the rule-based labels, KcBERT, and TF-IDF achieved accuracy/macro F1 values of 0.6762/0.6683, 0.6333/0.6240, and 0.5405/0.5068, respectively. These findings provide proof-of-concept evidence that the 7C framework can serve as a basis for organizing large-scale app-review data, while application of the trained classifier in the proposed operational monitoring phase remains to be empirically evaluated.