As artificial intelligence (AI) systems assume greater responsibility in educational assessment, questions surrounding fairness, transparency, and trust have become central to their ethical and pedagogical legitimacy. Yet, little empirical work has examined how specific design features shape students’ trust in AI-driven assessment, particularly in contexts where algorithmic decisions carry meaningful academic consequences. This study examines how transparency, ethical framing, and user agency influence students’ trust in an AI-based assessment platform. Using a 2 × 2 × 2 between-subjects experimental design with 240 undergraduate participants, the study isolates the main and interaction effects of these variables on trust, perceived fairness, perceived control, and adoption intention. Findings indicate that transparency is the most influential predictor of trust, while user agency functions as a compensatory mechanism in low-transparency conditions. Ethical framing, although theoretically salient, showed limited impact once users interacted with the system directly and shifted their attention toward the more concrete procedural cues embedded in the interface. A significant interaction between transparency and agency underscores the importance of aligning epistemic clarity with procedural control to foster behavioral commitment. These results support a multidimensional model of trust that incorporates emotional security, procedural justice, and behavioral intent. Overall, the study underscores that trust in AI assessment is not a byproduct of system accuracy alone but a reflection of students’ perceived legitimacy of the evaluative process.

The rapid emergence of generative artificial intelligence (GenAI) has introduced both opportunities and challenges for education systems worldwide. Educational stakeholders are grappling with fundamental questions of how to guide students on whether and when, and in what ways, they should use GenAI. In this study, a framework was developed to guide K–12 policies and guidelines on the use of GenAI. Using the Delphi technique and collective writing, expert perspectives were gathered from participants across 20 countries and six continents. The analysis identified eight key topic areas for K–12 GenAI policy and guideline development: (1) data privacy and security, (2) ethical and responsible use, (3) equitable access, (4) academic integrity, (5) human oversight, (6) GenAI literacy, (7) curriculum integration, and (8) governance and review. A complementary six-part framework was also constructed to support policy relevance and currency through multi-stakeholder governance, continuous review, ongoing training, awareness of external developments, outcome monitoring, and transparent communication. Together, these frameworks advance the scholarly and practical understanding of how GenAI policies can be designed and maintained in schools.

Developing highly active electrocatalysts for the oxygen evolution reaction is a pivotal challenge in sustainable water electrolysis. Herein, we report a novel in situ oxidative phase-restructuring strategy to fabricate oxygen vacancy-rich NiFe2O4 (NFO) directly on nickel foam. Distinct from conventional hydrothermal methods that typically yield thermodynamically stable crystals with limited intrinsic defects, our unique one-pot process involves the formation of a reduced metallic intermediate. The subsequent drastic phase transformation from this metallic state to a spinel oxide thermodynamically enforces the generation of abundant oxygen vacancies to relieve lattice stress, resulting in unique polycrystalline nanocrystal assemblies (NFO-1). Electrochemical evaluations reveal that NFO-1 significantly outperforms its thermodynamically equilibrated counterpart (NFO-2), exhibiting a low overpotential of 330 mV at 20 mA cm−2 and a remarkable mass activity of 6.78 A g−1. This superior performance is primarily attributed to intrinsic oxygen vacancies generated during the oxidative phase evolution, which optimize the active-site electronic structure and enhance charge–transfer kinetics. Furthermore, the catalyst demonstrates excellent durability over 1200 cycles. This work highlights oxidative phase restructuring as a powerful pathway to engineer intrinsic defects for high-efficiency energy-conversion applications.

Post-disaster reconstruction projects (PDRP) are integral to ensure that a community will recover and return to normal after a major disaster. The project team success (PTS) in PDRPs is essential to ensure that post-construction efforts will be effective and attain its objective of recovery in the community. An Artificial Neural Network (ANN) model was established considering several factors including post-disaster reconstruction project including project manager's leadership style (PMLS), multi-disciplinary project competence (MDPC), project manager's experience and competence (PMEC), high degree of trust within the project management team (HDTPMT), implementing an effective decision (IAED), effective project control (EPC), competent project manager (CPM), project risk and liability management (PRLM), motivated and well-integrated team (MWIT), and team composition (TC). The governing ANN model has a topology of 10-3-1 network structure and showed good performance with correlation plot (R) of 0.99850, MSE and MAPE of 0.00135 and 0.40559, respectively. The relative importance (RI) of the input parameters (IP) was also determined utilizing the connection weights (CWs) via Garson's algorithm (GA). The findings showed that the MWIT factor is the most influential factor (MIF) to project team success in PDRPs. The results in this study could be utilized to focus on improving areas to guarantee the success of PDRPs.

Steel diagrid high-rise structures require repeated finite-element analyses to accurately predict the multi-directional displacements, which is a time-consuming approach for parametric exploration and early-stage design. This paper presents an artificial neural network (ANN) – particle swarm optimization (PSO) informed model for predicting multi-directional displacements of high-rise steel diagrid frames considering different parameters including the number of storeys (NS), diagrid angle (DA), cross-sectional area (CSA), total weight (TW), and mass of the diagrid exterior (MDE). The model was developed from a dataset of 360 simulations from SAP 2000 ranging from 20-80 storeys and 33.69°-90° angles was used to create a Levenberg-Marquardt (LM) ANN with hyperbolic tangent sigmoid (HTS) activation function and 11 hidden neurons. The PSO was integrated into the model to enhance the training by optimizing the weights and biases (WB) of the network. The ANN-PSO achieved excellent model performance results with R values ranging from 0.9931 to 0.9989 and mean squared error (MSE) ranging from 0.000380 to 0.017200. The sensitivity analysis (SA) utilizing Garson's algorithm (GA) revealed that the number of storeys and diagrid angles are primary influencing the X and Y-displacements while the total weight and cross-sectional area were the leading influential factors to the Z-displacement. The proposed ANN-PSO offers an accurate, interpretable and computationally efficient approach for performance-based preliminary design of steel diagrid high-rise structures.

The metaverse and artificial intelligence (AI) are increasingly intersecting in educational contexts, yet limited empirical research has examined how generative AI avatars function within socially interactive virtual environments. This study investigates the deployment of generative AI avatars within a socio-educational metaverse environment. Using a developmental evaluation approach, data were collected through interviews with seven institutional stakeholders, teacher-generated reflections, internal documentation, embedded user feedback captured through in-platform reporting tools, and longitudinal field memos across an iterative deployment cycle. Findings indicate that the transition from scripted NPCs to generative AI avatars recalibrated users’ attribution of agency, intensified dialogic unpredictability, and elevated social realism beyond visual fidelity. Voice-mediated interaction emerged as a threshold mechanism for co-presence, while algorithmic improvisation exposed tensions between pedagogical intent and stochastic response generation. The deployment further revealed affective frictions, expectation misalignments, and the mediating role of AI literacy in shaping trust, participation, and interpretive coherence. Overall, the study advances a sociotechnical understanding of AI avatars as co-constructors of meaning and interaction, offering implications for the design, implementation, and governance of future AI-enhanced metaverse learning environments.

As artificial intelligence (AI) continues to permeate the education sector, the question is no longer about what it can do but what drives its successful adoption. Despite a growing body of literature on AI in education, research specifically addressing its adoption in developing countries is still lacking, even though the use of AI is potentially even more widely adopted there than in other places. Thus, we examined the critical factors influencing AI adoption in Ghanaian higher education institutions. Anchored in the diffusion of innovation (DoI) Theory and the unified theory of acceptance and use of technology 2 (UTAUT2), we specifically investigated the role of interoperability, relative advantage, pedagogical alignment, accessibility and affordability, and ethical considerations in shaping its adoption in these institutions. Quantitative data from 230 participants across 34 universities in Greater Accra was analyzed using a structural-equation-modeling approach. With an explanatory power of 77.3%, our model confirms the significant role of all five factors in shaping AI adoption. Our findings highlight the necessity for structured AI implementation strategies, including phased rollouts, professional development initiatives, and continuous system optimization to facilitate sustainable integration in resource-limited contexts. This study provides empirical evidence to guide policymakers and institutional leaders in aligning AI-driven educational innovations with strategic and contextual imperatives.

Brain tumor detection using magnetic resonance imaging (MRI) plays a critical role in early diagnosis and treatment planning. However, manual analysis of MRI images can be time-consuming and prone to human error. This study proposes a deep learning framework for brain MRI classification that integrates Contrast Limited Adaptive Histogram Equalization (CLAHE) preprocessing with a Multi-Stage Transfer Learning (MSTL) strategy. The proposed framework evaluates three convolutional neural network architectures, MobileNetV2, ResNet50, and EfficientNet-B0, to analyze their performance in classifying 3D-rendered brain MRI images into tumor categories. CLAHE was applied to enhance image contrast and improve the visibility of structural patterns before training. The MSTL framework progressively fine-tunes pretrained models through multiple stages, enabling better adaptation of learned features to the MRI dataset. Experimental results demonstrate that all three models achieved high classification performance. Among the evaluated architectures, ResNet50 achieved the highest accuracy of 99.06%, followed by EfficientNet-B0 at 98.90% and MobileNetV2 at 98.12%. Training curves and confusion matrix analysis further confirmed stable convergence and strong classification capability across the models. The novelty of this study lies in combining CLAHE-based MRI enhancement with a progressive transfer learning framework to improve deep learning performance in medical image classification. The proposed approach may support AI-assisted diagnostic systems for automated brain tumor detection and improve the efficiency of clinical decision-making processes.

Artificial intelligence (AI) has reshaped various aspects of human lives, particularly through its capabilities to address complex sustainability challenges. Despite the rapid expansion of AI applications, their contribution to cleaner production and sustainable development remains underexplored, especially in developing nations. In Southeast Asia (SEA), where AI adoption has grown substantially across environmental, economic, and social dimensions, research that examines its role in cleaner production outcomes remains fragmented. In view of this gap, this study conducts a systematic literature review (SLR) of AI applications related to cleaner production and sustainable development by examining relevant themes, application areas, and sustainability dimensions addressed by AI, while evaluating the maturity of AI methodologies, alignment with cleaner production outcomes, and integration with circular economy and resource efficiency goals. Moreover, it investigates the barriers and challenges that constrain AI application and offers future research directions to advance AI deployment for cleaner production and sustainable development across SEA countries.

As calls to center lived experience in higher education intensify, so too do concerns about the ethical, emotional, and structural risks involved in integrating real-life narratives into pedagogy. This study introduces Simulated Lived Experience (SLE) as a novel pedagogical modality that leverages the immersive affordances of learning environments like the metaverse to approximate systemic conditions of marginalization without reproducing trauma or requiring emotional labor from marginalized individuals. Drawing on critical pedagogy frameworks and affect theory, the research explores how SLE enables learners to engage with ethical discomfort, narrative complexity, and affective dissonance through the enactment of virtual selves. A qualitative design was employed, with data collected via semi-structured interviews from 12 participants who engaged in metaverse-based simulations portraying exclusionary dynamics related to disability, race, and institutional access. Thematic analysis generated four key findings: (1) virtual simulations evoke affective authenticity but also ethical unease, (2) embodied disorientation fosters structural insight, (3) narrative authorship and representation are ethically contested, and (4) discomfort acts as a catalyst for critical reflection. The study concludes that while SLE cannot replace lived experience, it can function as a powerful epistemic mediator when designed collaboratively, approached reflexively, and grounded in epistemic care.