The pervasive adoption of AI and AIoT applications at the network edge presents both opportunities and challenges for smart cities. With a focus on the energy efficiency of AI in urban environments, this paper provides a systematic comparative analysis of representative edge hardware platforms, i.e., embedded GPUs, FPGAs, and ultra-low-power microcontroller-/sensor-class devices, assessing their suitability for AI workloads in IoT-driven smart city infrastructures. The evaluation, based on direct characterization of diverse neural networks and relevant datasets, quantifies computational performance and energy behavior through inference latency, throughput, and energy/per inference measurements. Across the evaluated network–board pairs, the measured inference power spans several orders of magnitude, ranging from 0.1–10 mW for ultra-low-power Intelligent Sensor Processing Units (ISPUs) up to 1–10 W for embedded GPUs, highlighting the wide design space between the least and most power-demanding configurations. Results indicate that embedded GPUs provide a favorable performance-to-power ratio for computationally intensive workloads, while MCU/ISPU-class solutions, despite throughput limitations, offer compelling advantages in ultra-low-power scenarios when combined with quantization and pruning, making them well-suited for distributed sensing and actuation typical of smart city deployments. Overall, this comparative analysis guides hardware selection for heterogeneous, sustainable AI-enabled urban services.

Extending the UTAUT Model: The role of cognitive flexibility in AI adoption in higher education.

Artificial intelligence has emerged as a promising tool in fetal medicine, with applications in prenatal imaging, anomaly detection, and biometric analysis. Peer-reviewed studies have reported high accuracy for AI models in identifying congenital heart defects, segmenting brain structures, and predicting fetal growth patterns. Despite strong retrospective performance, most tools remain investigational due to limited external validation, lack of explainability, and poor integration with clinical workflows. This review synthesizes current evidence on AI applications in fetal diagnostics, highlights both capabilities and limitations, and outlines future directions needed for safe and effective clinical translation.

A RRA Perspective on AI and Machine Learning Applications in Radiology: From Experimental to Clinically Viable Solutions.

Hyper-dimensional computing architectures for AI applications: A comprehensive survey of principles, performance, and future prospects

Prompt engineering in generative AI applications for commercial single pilot operations (SiPO): An emerging competency

The increasing complexity of tax systems and the limitations of traditional rule-based audits have highlighted the need for adaptive, transparent, and efficient auditing solutions. This paper presents the design and evaluation of a Secure AI-Driven Adaptive Audit Transparency Engine (AI-AATE), a novel framework integrating machine learning, explainable AI (XAI), and human-in-the-loop oversight to enhance tax compliance, reduce administrative inefficiencies, and strengthen economic outcomes. The architecture combines supervised and unsupervised models for risk detection, continuous feedback incorporation for adaptive learning, and comprehensive audit logging to ensure transparency, fairness, and traceability. A rigorous evaluation framework employing operational Key Performance Indicators (KPIs), counterfactual simulations, and economic modeling quantifies performance across audit yield, coverage, processing efficiency, revenue recovery, and equity. Governance and trust metrics assess explainability, human oversight, and bias mitigation, linking design principles to measurable institutional outcomes. Simulation results demonstrate that AI-AATE can significantly improve detection of non-compliance, optimize resource allocation, and support equitable and accountable audit selection compared to traditional approaches. By bridging technical design, performance evaluation, and economic impact assessment, this study contributes a holistic methodology for AI-enabled audit systems, offering actionable insights for policymakers, tax authorities, and researchers. The findings underscore the potential of AI-AATE to transform public-sector auditing while maintaining fairness, legitimacy, and public trust, addressing a critical gap in the literature on adaptive, transparent, and secure AI applications in taxation.

This article aims to provide an overview of the potential uses of AI in the military and to emphasize the need to identify and define measurable indicators in order to evaluate the benefits of cutting-edge technologies and solutions that are expected to improve the quality and performance of operations. It focuses on crucial areas such as situational awareness and decisionmaking support, as well as logistical and operational planning and modelling and simulation (M&S). AI is becoming a crucial tool for intelligence and intelligence analysis of the enemy, and its role in military operations planning and support is growing. Autonomous vehicle and weapon systems are another area in which AI can be put to use. The use of AI is expected to have a greater impact on the military functions of human-machine interfaces (machine-learning, man-machine teaming). AI promises to overcome Big Data's "3V challenge" (volume, variety, and velocity), as well as the "2V challenge" (veracity, value) and render data processing at a controlled level of decision-making based on AI's knowledge. In this essay, we will talk about a number of AI applications in the military, their capabilities, opportunities, and the harm and destruction they could cause in times of instability. The seven AI patterns, the military's use of AI algorithms, object detection, military logistics, and robots, the global instability caused by AI use, and nuclear risk constituted the majority of the discussion.

Abstract This study investigates the impact of AI-assisted blended instruction on EFL learners’ speaking performance and learning resilience. Grounded in Social Cognitive Theory, the Technology Acceptance Model, and Sociocultural Theory, the research addresses a growing need to understand how AI tools shape not only linguistic accuracy but also learners’ psychological adaptability. Adopting a quasi-experimental design, the study compared an experimental group using ChatGPT and Gemini within a task-based learning framework to a control group receiving conventional instruction over an 18-week period. Quantitative analyses revealed significant gains in pronunciation and accuracy among the experimental group, with coherence showing marginal improvement and fluency remaining unchanged. In terms of resilience, learners demonstrated marked increases in metacognitive and social resilience, whereas ego resilience remained unaffected. These findings suggest that while AI tools can effectively support linguistic development and foster strategic learning behaviors, they may be less effective in cultivating emotional adaptability. The differentiated outcomes highlight the importance of aligning specific AI functionalities with targeted pedagogical goals rather than assuming uniform benefits. This study contributes a nuanced perspective to the discourse on AI in language education by uncovering domain-specific effects and clarifying the boundaries of AI’s pedagogical influence. It calls for future research to explore the long-term impacts of AI integration on learner autonomy, motivation, and affective growth, thereby paving the way for more intentional, theory-driven applications of AI in EFL contexts.

This paper discusses the role of artificial intelligence and its impact on human resource management outside of business in health service provider such as a public hospital. The aim is to investigate the results of the operation of artificial intelligence in the health sector and its use in human resource management in a public hospital, through applications that are already used in the business world. In the first phase in AI’s before and after the implementation of an online human resource service platform in a public hospital and a second bibliographic research was chosen to present and analyze specific scenarios of the application of AI’s by the business world, as good practices. The comparative evaluation resulted in an increase in use of 18.66%, a reduction in errors of 3600%, a reduction in service time of 271%, increased user satisfaction of 144% and a reduction in staff employed of 20%. The use of AI’s applications in human resource management is perhaps the critical element for the effective and efficient operation of a hospital in human resource management and the satisfaction of employees and patients although the research is limited to a single hospital and the existence of additional AI applications in Human Resource Management is a limitation.

As generative AI applications in supply chain management become increasingly thorough, systematic studies on how it could promote enterprise innovation are yet to come to light. This paper takes 298 manufacturing enterprises in Zhejiang Province as samples, uses questionnaire surveys and PLS-SEM methods to investigate how generative AI exerts its influence on supply chain innovation, and tests the role of knowledge sharing and supply chain learning as a mediator. Research has found that generative AI capabilities can significantly enhance knowledge sharing and supply chain learning levels. Knowledge sharing not only promotes supply chain learning but also has a direct driving effect on supply chain innovation, playing a key mediating role between generative AI capabilities and innovation. In contrast, the hypothesised mediation of supply chain learning did not receive statistical support. This indicates that the impact of generative AI on supply chain innovation does not depend on supply chain learning. The results reveal the transmission path of generative AI in supply chain innovation, emphasising the core position of knowledge sharing in the process of transforming technological capabilities into innovative results. This paper provides new empirical evidence to understand AI-driven innovation and provides reference practice to promote digital transformation and collaborative innovation among manufacturing enterprises.

Abstract Global healthcare organizations have been rapidly transitioning to digital use including electronic health records, telemedicine platforms, AI applications, and advanced digital communication technologies (Hanelt et al., 2021). While these new technology-driven approaches serve to both increase operational efficiency and clinical decision-making and improve patient outcomes, they generate technical demands driving technostress among healthcare providers (Ragu-Nathan et al., 2008; Tarafdar et al., 2007). Technostress has been correlated with reduced employee well-being, decreased performance and reduced innovative work behaviour (Ayyagari et al., 2011). In this regard, digital leadership is seen as a key organizational capability that can reduce this technological stress in the workers and enable innovation and adaptability of staff (Claassen et al., 2021; Cortellazzo et al., 2019). This article provides a conceptual synthesis on technostress, digital leadership, innovation climate and innovative work behaviour in a context of digitalization of healthcare. Based upon Job Demands–Resources theory (Bakker & Demerouti, 2017), Conservation of Resources theory (Hobfoll, 1989), Social Exchange Theory (Blau, 1964), and Organizational Climate Theory (Amabile, 1996; Hülsheger et al., 2009), the present study proposes that digital leadership mediates between technostress and innovative work behaviour and innovation climate moderates the effect. The paper informs practice by contributing theoretically to assimilation while offering implications for healthcare leadership in the field from which it can draw future empirical directions to research in such digitally transforming healthcare settings. Keywords: Technostress; Digital Leadership; Innovation Climate; Healthcare Digitalization; Innovative Work Behaviour; Organizational Innovation.

AI-enabled Business Intelligence (BI) solutions provide dramatic improvements in automating regulatory reporting, accuracy, efficiency, and compliance. The present paper investigates applications of AI technology, like machine learning and natural language processing, to simplify collecting, validating, and reporting regulatory data. Through automation, AI applications lower human effort, minimize errors, and deliver timely, compliant submissions. The analysis reviews a range of BI systems and how effective each has been in streamlining report automation, focusing on compliance rates and operational efficiency improvements. The primary findings are that AI-enabled solutions can cut reporting time and errors by a significant proportion, as well as overall compliance rates. The document also addresses issues of implementing these programs, such as integrating data, managing system compatibility, and resistance from companies. The results outline the potential of using AI to revolutionize regulatory reports, providing a scalable framework that enhances both efficiency and compliance across business sectors.

With the development of the digital intelligence era, generative AI is being widely used in scientific research, and its impact on graduate students' research competence has attracted much attention from the academic community. Based on cognitive distribution theory and self-efficacy theory, this study classifies AI applications into three levels from basic to advanced-technical support AI use, text development AI use, and transformation AI use-explores their effects on graduate students' research competence, and examines the mediating effect of critical thinking and the moderating effect of research self-efficacy. The results of the empirical analysis show that all three types of AI use behaviors are significantly correlated with research competence, with the strongest correlation for text development type and the weakest for technical support type. In the relationship between the three types of AI use behaviors and research competence, critical thinking plays a significant positive mediating role, and research self-efficacy plays a significant moderating role. Universities and tutors should guide students to focus on higher-order AI use behaviors in the text development and transformation categories, promoting the use of critical thinking to avoid technology misuse and improving research self-efficacy to help students accumulate confidence and support their research.

The inference of unstructured text semantics is a crucial preprocessing task for NLP and AI applications. Word sense disambiguation and entity linking tasks resolve ambiguous terms within unstructured text corpora to senses from a predefined knowledge source. Wikipedia has been one of the most popular sources due to its completeness, high link density, and multi-language support. In the context of chatbot-mediated consumption of information in recent years through implicit disambiguation and semantic representations in LLMs, Wikipedia remains an invaluable source and reference point. This survey covers methodologies for entity linking with Wikipedia, including early systems based on hyperlink statistics and semantic relatedness, methods using graph inference problem formalizations and graph label propagation algorithms, neural and contextual methods based on sense embeddings and transformers, and multimodal, cross-lingual, and cross-domain settings. Moreover, we cover semantic annotation workflows that facilitate the scaled-up use of Wikipedia-centric entity linking. We also provide an overview of the available datasets and evaluation measures. We discuss challenges such as partial coverage, NIL concepts, the level of sense definition, combining WSD and large-scale language models, as well as the complementary use of Wikidata.

With the rapid advancement of generative artificial intelligence (Gen AI) technology, an increasing number of studies are integrating Gen AI into healthcare. This study analyzed 1987 English publications in this field using bibliometric methods, sourced from the Web of Science Core Collection (WOSCC). The findings reveal a significant increase in publications since 2023, with 496 publications in 2023 and 1478 publications in 2024. The most contributing and influential journal was the Journal of Medical Internet Research. The total number of publications (TP) of this journal was 66, and the total number of citations (TC) was 1108. The most contributing country/region, affiliation, and author were the United States of America (TP=841, TC=8740), Harvard University (TP=89, TC=815), and Lechien, Jerome R. (TP=18, TC=228), respectively. The closest partnerships were observed between the USA and China, Tel Aviv University and Chaim Sheba Medical Center, and Cheungpasitporn, Wisit, and Thongprayoon, Charat, respectively. Research topics of all publications mainly focused on the application of Gen AI in clinical diagnosis, decision support, medical education, patient education, and mental health management, while also emphasizing technical and ethical challenges. Notably, several clusters highlighted the relevance of Gen AI in surgery, underscoring its potential impact in this key branch of healthcare. The findings will provide academic insights for technology developers and policymakers, as well as guidance for future research directions.

The adoption of the NEP 2020 highlighted the importance of teacher education and the application of AI in teacher education. In this research, researcher explore the apply of AI in teacher education in terms of opportunities, challenges, and equity. The researcher use a descriptive quantitative study design guided by a structured questionnaire for a sample of 60 teacher educators and pre-service teachers. The study survey elicited teacher educators’ and pre-service teachers’ understanding, perception, and attitude toward the use of AI in the ‘teaching-learning process,’ ethical issues, and equity. The outcome of the study shows that AI has the potential to enhance personalized learning and the development of skills in teacher education. However, concerns regarding privacy, bias, digital equity, and access remain. This study shows that AI, Ethical, and 2020 NEP capacity-building frameworks remain a desideratum.

Generative AI (GenAI) has attracted a surge of attention from higher education constituents after OpenAI released ChatGPT in November 2022. While numerous articles discuss applications and perceptions of GenAI in higher education, no comprehensive review has considered commonalities and differences among various educational stakeholder groups and contexts. In this review, we synthesize the applications, capabilities, and perceptions of GenAI in higher education to provide stakeholders (i.e., students, instructors, researchers, staff, and administrators) with insights into this topic to facilitate GenAI integration in higher education. We reviewed 50 relevant empirical articles published from January 2023 to April 2025 on GenAI in higher education. Our findings demonstrate how GenAI has already been applied and present its potential for implementation across teaching and learning, research, and student affairs in higher education. Among various stakeholders in higher education, students hold a more open and positive attitude toward this rising technology, while instructors and researchers hold mixed attitudes toward GenAI usage, and administrators tend to hold an open but cautious attitude toward GenAI implementation. Addressing common stakeholder concerns and needs, we outline institutional strategies for responsible GenAI integration, including launching GenAI learning hubs, formalizing license agreements, redefining academic originality, and implementing pilot programs.

Generative Artificial Intelligence (GenAI) poses major challenges for the field of education and so-called big language models are increasingly being applied in the work of education both in their pedagogical training and in their teaching. In our paper, we analyze how AI can be used to support the achievement of different types of cognitive goals defined within the framework of Bloom's taxonomy, which has been revised several times, while offering educators and researchers an overview of useful prompts that can be effectively used to support teaching, as well as to support students' classroom or home preparation (we specifically focus on the use of AI in language education). In doing so, we draw on concepts defined by researchers at Oregon State University, the SAMR model, and we also leverage our experience with the development and use of the Khanmigo AI application. We address both simple and complex AI prompt creation, the creation of personalized AI assistants, AI-enabled gamification, and other ways AI can be used to effectively support learning. We also focus on the requirements for constructing functional prompts – minimizing hallucinations or biases.

ABSTRACT Since ChatGPT’s public release in late 2022, discussion in marketing education has focused on applications of AI in teaching and learning. Less attention, however, has been given to the essential role that wellestablished marketing theory and sound pedagogy play in guiding these applications. This paper introduces the LEO Framework – a pedagogical model to help business educators integrate AI into course design in ways that advance student learning. The conceptual framework connects disciplinespecific theory with educators’ preferred pedagogical approaches, enabling them to align AI use with clear educational goals and equip students to learn more deeply in an AI-enhanced environment.