Purpose While Artificial Intelligence (AI) is increasingly embedded in Open Government Data (OGD) ecosystems, the mechanisms through which it restructures information practices and governance remain insufficiently theorized. This study empirically evaluates and refines an AI–OGD typology, explaining how AI roles are structured, sequenced and shaped by contextual conditions. Design/methodology/approach A mixed-methods expert panel evaluation (N = 9) was conducted with specialists in information systems and public administration. Five structured tasks were used to assess the conceptual clarity, role distinctiveness, sequencing and strategic positioning of the proposed AI roles, combining qualitative assessment with quantitative measures of expert agreement and perceived influence and impact. Findings Results support the typology while extending its theoretical interpretation. Rather than a flat classification, AI roles are perceived as forming a hierarchical and interdependent system, progressing from foundational portal-oriented functions to high-impact ecosystem-level roles. The findings further reveal low classification accuracy as an indicator of functional overlap, suggesting that AI roles operate along a continuum rather than as discrete categories. In addition, implementation patterns are shown to be cyclical and contingent on socio-technical factors such as organizational readiness, data infrastructure maturity and governance capacity. Practical implications The study provides a structured roadmap for policymakers and public managers to sequence AI adoption, balancing foundational data capabilities with ecosystem-level value creation. Originality/value This study contributes beyond validation by reconceptualizing AI roles in OGD ecosystems as a dynamic governance architecture. It advances research by integrating empirical validation with explanatory insight, offering a more practice-sensitive and theoretically grounded understanding of AI-enabled public data ecosystems.
Purpose This study investigates linguistic anthropomorphism in large language models (LLMs) by analyzing over one million conversational turns generated by 25 state-of-the-art models, including GPT-4, Claude, Vicuna, Llama-2, and others. The research examines anthropomorphic language markers across eight major languages and explores variation in anthropomorphic language use by model and language. Design/methodology/approach Using the Large Model Systems (LMSYS) Chat 1M dataset from Hugging Face, the study examines anthropomorphic language markers such as first-person pronouns, explicit empathy statements, and emojis across eight major languages. The methodology combines the detection of predefined anthropomorphic and empathetic phrases, the quantification of first-person references, and sentiment analysis to explore variation in anthropomorphic language use by model and language. Findings Results reveal substantial differences in the prevalence of anthropomorphic expressions and affective tone across LLMs, with certain models displaying significantly higher rates of empathetic and self-referential language. The analysis further identifies linguistic and cultural effects, showing that the propensity for anthropomorphic responses is not uniformly distributed across languages or cultural contexts. These findings support the theoretical concept of Situated Linguistic Anthropomorphism, according to which human-like AI language emerges from the interaction between model design and communicative setting. Originality/value The findings provide new insights into the design and evaluation of conversational AI, emphasizing the relevance of anthropomorphic cues for user trust and engagement in multilingual settings. The study contributes to understanding how linguistic anthropomorphism varies across AI models and cultural contexts, with implications for responsible AI design. Peer review The peer review history for this article is available at: Link to the website
Purpose The rank of a journal is usually quantified by a single metric like Impact Factor (IF), CiteScore or SCImago Journal Rank (SJR). To avoid bias due to the use of a single metric, a multimetric-based data-centric unsupervised machine learning framework is proposed for journal categorization. Design/methodology/approach This study integrates six bibliometric indicators as features to map the latent structure of journal impact. The proposed framework incorporates two state-of-the-art dimensionality reduction algorithms (PCA and t-SNE), which also contribute to feature engineering. Using PCA, the correlated metrics are transformed into reduced dimensions (2D/3D) that serve as input for k-means clustering. The optimal clusters (k = 4), aligned with IF, Scopus and SJR quartiles, are validated through internal and external evaluation measures. Findings Multimetric-based clustering of 4,146 computer science (CS) journals is compared with established single-metric approaches (SJR, IF and Scopus quartiles) and with HJRS's proprietary multimetric approach, adopted in Pakistan. The average metric-based results provide interpretable outcomes against SJR, IF and Scopus quartiles (single-metric) as well as HJRS (multi-metric). The proposed journal categorization method is fully automated and offers improved classification of journals compared to well-known traditional metrics. Originality/value The contribution of this study is evident in the proposed framework for categorizing journals using multiple metrics. The novelty of our work lies in the integration and domain-specific adaptation of unsupervised machine learning to categorize CS journals. Peer review The peer review history for this article is available at: Link to the website
Purpose This study aims to assess the current state of digital literacy (DL) among Vietnamese university students in the context of digital transformation. The findings provide a foundation for recommending strategies for higher education institutions and academic libraries to enhance students' DL. Design/methodology/approach The study employed a self-assessment survey with 1,154 students selected through convenience sampling combined with purposive elements. The measurement scale was developed based on UNESCO's Digital Literacy Framework (2018) and the Jisc Digital Capabilities Framework (2024). Data were analysed using descriptive statistics. Findings Results show that students' DL is at a high level, with the highest scores observed in basic skills such as device-software operation, information handling, and digital communication. More advanced skills – including digital content creation, digital learning skills, and career-related digital competencies – were rated lower. Male students reported higher scores in technical skills, whereas female students reported higher scores in creativity and learning-related skills. Students in Education and Pedagogy reported the highest levels of DL. Third-year students reported the highest DL scores. DL scores differed across daily Internet-use groups, but the pattern was not linear. Originality/value This study provides large-scale empirical evidence on DL among Vietnamese university students, integrating internationally recognised frameworks adapted to the local educational context. The findings highlight the multidimensional nature of DL and underscore the critical role of higher education institutions and academic libraries in fostering targeted digital skill development for students. Peer review The peer review history for this article is available at: Link to the website
Purpose Facebook support groups provide opportunities for emotional support and social connection among individuals with shared health conditions. This study examined the relationships among online engagement, perceived social support, self-compassion and well-being among adults with attention difficulties (ADHD) participating in Facebook groups. Design/methodology/approach The sample included 405 adults who reported having ADHD or attention difficulties, recruited from two types of Facebook groups: members of ADHD-specific Facebook support groups and members of unrelated, neutral groups. Four key variables were assessed: online engagement, social support, self-compassion and well-being. Analyses included descriptive, comparative, correlational, regression and structural equation modeling techniques. Findings Significant differences between the types of groups were found in two of the four variables: self-compassion and well-being, with both showing significant interactions with gender. Women in ADHD-related groups reported lower self-criticism than women in neutral groups, suggesting that emotional support in these groups may help reduce self-criticism. Conversely, men in ADHD-related groups reported lower levels of well-being than in neutral groups, indicating that they may experience less comfort engaging in emotionally supportive interactions online. Originality/value These results highlight the potential of Facebook groups to function as support systems that enhance well-being for individuals with attention difficulties. They also raise important questions regarding the role of gender in shaping experiences of belonging and support in digital environments.
