This paper explores an emerging dimension of AI-enhanced assessment practice - machine metacognition - through a case study of an adaptive coaching application built to operationalize the "design and systems thinking for transformation" (DOT) Framework. Traditionally, assessment has focused on evaluating human learning or performance, sometimes with the assistance of machines (e.g., automated scoring and, more recently, AI), while system performance of AI has been evaluated externally by human observers. In contrast, this study focused on an iterative process of human-prompted 'machine self-assessments' of the monitoring and control of the reasoning, responsiveness, and progression of an AI coach during cycles of test sessions with users. We analyze transcripts from a set of five cases to document the extent to which, and how, the Coach recognized, evaluated, and adjusted its actions across the stages of a variety of frameworks provided as contexts for generating coaching advice. Findings highlight evidence of what we have termed 'machine metacognition,' instances when the system, whether prompted or not, recognized misalignment, corrected course, or signaled uncertainty. We demonstrate how such contexts and mechanisms may redefine the boundaries of AI-enhanced assessment. The paper argues that recursive self-assessment within intelligent tools like the AI coach signals a conceptual shift toward double-loop assessment, in which both human and machine continuously learn about and from their evaluative processes.
Students' classroom interest is a critical factor influencing learning outcomes, and its intelligent detection has become an important issue in smart education. Traditional assessment relies on teacher observation or post-class questionnaires, which are time-consuming, labor-intensive, and subjective, and which lack real-time feedback. Some studies infer interest using physiological signals from wearable EEG or ECG devices, but these methods may interfere with normal learning. The study focuses on the typical classroom teaching scene and proposes a classroom interest assessment method, which is designed to achieve non-intrusive, contactless, real-time feedback using computer vision techniques. The proposed method extracts three types of visual cues from students during classroom learning: behavior, attention, and facial emotion. Each dimension is quantified to generate a corresponding score, and the three scores are integrated to assess students' classroom interest levels. Experiments on a self-constructed classroom dataset show that the proposed model can effectively recognize classroom interest levels.
This study investigates the effectiveness of a computer-based learning environment grounded in the Cognitive Theory of Multimedia Learning (CTML) for improving middle school students' conceptual understanding of optics and inquiry skills in two distinct contexts: Turkey and Indonesia. Using a quasi-experimental design with 134 eighth-graders, participants were assigned to either a CTML-informed e-learning experimental group or a traditional instruction control group. Data from two-tier conceptual understanding and inquiry skills tests were analyzed using non-parametric methods (Wilcoxon signed-rank and Mann-Whitney U). Results indicated the experimental group achieved significantly greater improvements in both variables compared to the control. Notably, cross-cultural comparisons revealed a significant divergence in post-intervention conceptual understanding favoring Turkish students, despite both groups exhibiting statistically similar inquiry skills and baseline knowledge. Medium-to-large effect sizes support the efficacy of CTML principles in realistic settings. This research addresses gaps in scientific inquiry for underrepresented middle school populations and demonstrates that a comprehensive suite of multimedia principles can optimize learning in complex, authentic environments. The findings enhance the generalizability of CTML in diverse educational systems, particularly in low- to middle-income country settings.
Social media and its continuous influx of information and content have become a part of everyday reality for K-12 students. Educators face the critical challenge of equipping their students with the necessary digital citizenship knowledge and skills to protect against the ever-evolving online threats to their safety, privacy, and well-being. However, digital citizenship currently lacks a universal definition and a standardized skillset. This paper presents a Delphi study designed to identify the essential elements for a K-12 digital citizenship curriculum. The study implemented three iteratively designed surveys to collect opinions from twenty-two expert panelists regarding key digital citizenship knowledge and skill items for inclusion in K-12 curricula. The study resulted in a list of twenty items, which includes six newly generated items that address topics such as AI use, and a new category, Knowledge vs. Behaviors, which was not included in the previous version of the Nine Elements of Digital Citizenship.
With the rapid development of educational technology, Virtual Reality (VR) is increasingly applied in classroom teaching reform. Based on the Dynamic Interaction Five-Stage Instructional Model (DIFI model), this study uses Lag Sequential Analysis (LSA) to systematically compare teacher-student interaction behaviors in VR and traditional classrooms, and to examine the DIFI model's role in optimizing interaction structure in VR settings. Eight teaching videos across primary and junior secondary subjects were selected. By constructing a teacher-student interaction coding system, the study analyzes the frequency, sequential patterns, and technological involvement in interactions. Results show that in DIFI-based VR classrooms, a dynamic balance between teacher guidance and student initiative is achieved, and technology is deeply integrated, enhancing interaction frequency and complexity. Behavioral sequences are more diverse, and student participation is higher than in traditional classrooms. These findings broaden perspectives on VR classroom interactions and offer theoretical and practical guidance for future virtual teaching design.
This study examines how two Norwegian grade 8 students described their use of mathematical problem-solving heuristics and computational thinking (CT) concepts, CT practices, and CT perspectives when engaging with interdisciplinary problem-solving tasks. Under the qualitative case study design, data were collected through semi-structured recall interviews conducted after the students had completed two PISA problem-solving tasks. The analysis focused on the two students' problem-solving strategies, applying frameworks from mathematical problem solving and CT alongside horizontal and vertical transfer of knowledge. The findings show that both students activated relevant knowledge from mathematics and CT but differed markedly in how they coordinated, monitored, and adapted these strategies. One student demonstrated flexible integration and evidence of vertical transfer, while the other applied CT concepts procedurally, resulting primarily in horizontal transfer. These contrasting patterns highlight how strategic control shapes interdisciplinary problem solving and underscore that productive integration requires opportunities for students to evaluate representations, justify their strategic choices, and adapt their reasoning across contexts. The study concludes that educators can support vertical transfer by designing tasks and scaffolds that make structural relationships explicit, prompt students to monitor and revise strategies, and help them link CT practices with underlying mathematical concepts.
The development of coding skills is considered important for a digital age, linked to broader learning and non-cognitive abilities. Schools in England integrate coding into the curriculum, perhaps supplemented by after school activities. "Code Club" is a structured learning programme offered as an after-school activity for pupils aged 9 to 13. This study evaluates Code Club using a quasi-experimental design assessing the impact on student attitudes to learning and coding skills. Pupils attending the 22-week programme were compared to non-participants in the same schools. Initially, 412 pupils (Years 4-9) participated, but only 239 from 13 schools remained for full analysis. The findings suggest positive effects for resilience (+0.22), confidence (+0.47), sense of belonging (+0.09), and coding skills (+0.24). 40% dropout rate limits the generality of the conclusions. Code Club, run voluntarily in schools, was well-received, with participants enjoying coding and digital projects. The paper discusses the findings and implications.
This qualitative interpretive study examined how an AI-assisted mobile application supported a differentiated professional development opportunity (DPDO) for 30 STEM teachers across six Arab countries over 24 weeks. Grounded in andragogy and interpreted through a TPACK lens, the study explored teachers' perceived benefits of using the SchooTeach mobile app for AI-aware professional learning. Post-program focus group interviews were thematically analyzed, yielding four interrelated areas: the differentiated professional development program, the mobile application, AI in STEM education, and the virtual community of practice. Across these areas, teachers reported that the DPDO fostered agency through choice and ownership, collaborative learning ecologies, holistic professional identity, and metacognitive reflection, alongside tensions related to feeling overwhelmed and the need for clearer guidance. The app's intuitive interface, multimodal resources, and AI-supported recommendation engine enabled personalized, continuous learning, yet limited social interaction features constrained deeper collaboration. AI was described as expanding teachers' repertoire of STEM pedagogies and enhancing confidence, while raising concerns about data privacy, bias, and equity. Overall, the findings suggest that AI-aware, mobile-mediated professional development can strengthen teachers' AI-related technological knowledge and emerging TPK/TPACK without displacing human judgment, provided that designs scaffold choice, embed collaboration, and address ethical considerations.
This study investigated the effectiveness of AI-generated animated educational films in enhancing kindergarten children's first aid knowledge in Kuwait. A mixed-methods, quasi-experimental design was employed with 65 children assigned to an experimental group (n = 32) and a control group (n = 33). The AI-generated films were developed using generative AI tools to address the lack of Arabic-language, developmentally appropriate first aid resources for preschoolers and to produce culturally relevant, age-suitable content tailored to Kuwaiti children's real-world contexts. The experimental group received six sessions over two weeks covering nosebleeds, heatstroke, minor burns, sprains, bleeding, and the emergency number '112.' A pictorial cognitive test was administered as a pretest and post-test for both groups, with a delayed post-test for the experimental group after three weeks. Results showed significant improvement for the experimental group, whose mean score increased from 3.71 to 10.85 out of 12, while the control group showed minimal change. The intervention explained 97.6% of the variance (eta 2 = 0.9763) with a large effect size (d = 10.88). Knowledge retention was confirmed by the delayed post-test, further supported by qualitative observations indicating high engagement and accurate recall. Interpreted through Social Learning Theory, Cultivation Theory, and the Sociological Theory of Digital Learning, these results underscore the importance of behavioral modeling, repeated exposure, and interactive media for young children's learning. The study concludes that AI-generated animated films are an effective tool for early childhood health and safety education and recommends equipping kindergartens with digital resources, training teachers in multimedia integration, and involving specialists to ensure age-appropriate content. These findings highlight the potential of innovative AI-supported strategies for foundational health education in early childhood settings.
This qualitative case study examines how upper elementary educators integrate artificial intelligence (AI) tools into instruction through a sociocultural lens that foregrounds developmental appropriateness, ethical reasoning, and cognitive agency. Drawing on professional development field notes, exit tickets, and semi-structured interviews with eight teachers across grades 6-8, the study identifies three interconnected themes that reflect how teachers act as developmental mediators in the classroom: (1) Developmental Responsiveness and Student Variability in AI Integration, highlighting how teachers assess and respond to students' cognitive, behavioral, and ethical readiness; (2) Systematic Scaffolding for AI Engagement, detailing instructional strategies that support language complexity, executive functioning, prompt engineering, and the writing process; and (3) Critical AI Literacy and Ethical Integration, which centers metacognitive reflection, digital citizenship, and responsible AI use. Across all themes, teachers emphasized preserving student ownership, resisting over-reliance on automation, and fostering ethical deliberation through modeling, guided practice, and open dialogue. By treating AI as a cultural and cognitive tool, teachers supported students' critical engagement and moral development, ensuring that technology use aligned with their developmental trajectories. Findings underscore the need for sustained, developmentally informed professional learning and policy structures that support ethical and reflective AI integration in elementary settings.
The integration of Artificial Intelligence (AI) into K-12 education holds significant promise for transforming teaching and learning processes. Central to this transformation are teachers' perspectives, which play a key role in shaping the adoption and effective use of AI technologies in classrooms. Existing literature has not systematically synthesized these perspectives through the lens of well-established theoretical frameworks, nor developed structured models to guide evaluation efforts. To address these gaps, this study applies constructs from the UTAUT and TPACK frameworks to guide a thematic synthesis of findings, with the aim of identifying enabling and constraining factors influencing AI integration in K-12 education from the standpoint of educators. The review process followed PRISMA guidelines to ensure rigorous and systematic literature selection and analysis. Findings suggest that teachers generally exhibit a blend of optimism and cautious concern regarding the adoption of AI in K-12 education. We identified three critical factors as particularly influential: teachers' Technological Pedagogical Content Knowledge (TPACK), teacher agency, and teacher affective orientations. In response to the complexities of implementation, we propose a novel preliminary assessment model, guided by the evaluative principles of UNESCO's Global Education Monitoring Report 2023. The proposed model offers a practice-grounded application of global policy dimensions: Equity, Sustainability, Appropriateness, and Scalability, linked with synthesized classroom-level insights. It further delineates ten essential subthemes, providing a structured approach for evaluating the effectiveness of AI integration in educational settings.
Effective learning support systems are essential to ensure that all students, regardless of their individual abilities and learning needs, can achieve academic success. This study examines the implementation of support mechanisms in primary education, focusing on differentiated teaching strategies, digital tools, and individualized assistance within a competency-based learning framework. Using qualitative action research, including classroom observations and teacher interviews, the study analyzes how educators apply various support strategies to address diverse student needs. A convenience sample was used, consisting of 10 third-grade teachers who participated in the study conducted from May to September 2024, as competency-based education was newly introduced in the third grade that year. The findings reveal that structured learning support consists of multiple interconnected elements: digital platforms that allow for personalized learning at an individual pace, targeted teacher assistance that fosters student progress, and peer collaboration that enhances cooperative learning. Additionally, teachers implement flexible assessment methods that enable students to demonstrate their acquired knowledge in various ways. The results emphasize the need to develop a comprehensive and adaptable support system that integrates digital resources, teacher guidance, and collaborative learning structures. This study highlights the importance of continuous professional development for teachers to ensure that all students receive the necessary support, develop their competencies, and achieve successful learning outcomes.
Computational thinking (CT) has been perceived as an essential skill for everyone. Several countries and institutions have attempted to concretize CT in the schools' curricula. This paper presents the integration of CT into mathematics (Math + CT) lessons utilizing specific mathematics software, GeoGebra. Through the educational design research (EDR), we innovated our Math + CT lessons by integrating ChatGPT to generate GeoGebra commands. Five adults and 16 junior high school students participated in this study. We collected the data from participants' ChatGPT prompts and screen video recordings. The data were mainly analyzed through a content analysis method. The findings revealed that ChatGPT has limitations in generating GeoGebra commands but provided opportunities to develop students' CT skills. We also propose a model to inform educators and researchers about ChatGPT's strengths and weaknesses in supporting students' CT skills in GeoGebra, and to guide curriculum developers on integrating AI literacy and CT into mathematics curricula.
This paper explored in-service teachers' perceptions and approaches to computational thinking (CT) and block-based coding via Scratch before and after participating in a week-long online and asynchronous learning module within a graduate-level instructional technology course. We analyzed pre- and post-learning module data from 21 teachers across different grade levels and subjects and examined their approaches to CT, changes in their CT-related knowledge and skills, and views on the utility of Scratch. Our findings suggested a shift from pre-module findings, showing a change in their disposition toward Scratch's utility value and recognition of more subject-specific applications of Scratch, such as systems modeling in science and digital storytelling. In light of the findings, we highlight the recurring need for teacher education programs to support teachers' self-efficacy in teaching CT, the development of CT-related knowledge and skills, especially advanced coding practices, and to embed self-reflective questions that help teachers recognize CT's relevance to their disciplinary teaching areas and students' learning. We also offer implications for designing future teacher education programs.
This systematic review explores the potential of virtual reality-(VR)-based interventions in preparatory-level STEM education, incorporating Strengths, Weaknesses, Opportunities, and Threats (SWOT) analysis for a comprehensive evaluation. Following PRISMA guidelines, 35 articles published between 2000 and 2024 were selected based on predefined criteria, encompassing diverse international contexts. Analysis revealed that science predominantly leverages VR interventions at this level (n = 26). Game-based learning emerged as the most common pedagogical approach (n = 19), followed by inquiry-based models (n = 6). Key student outcomes targeted by VR-based instruction included cognitive (n = 25), affective (n = 24), and behavioral gains (n = 19). The SWOT analysis highlights strengths, weaknesses, opportunities, and threats, offering valuable insights for enhancing STEM education. This research informs policymakers on future VR initiatives and equips educators with strategies to create impactful, engaging VR-driven learning experiences, fostering meaningful advancements in preparatory-level STEM education.
Modern digital tools facilitate mathematics teachers to integrate new approaches to teaching and learning into classrooms and also to deal with additional, more sophisticated mathematical content in lessons. The intended forms of student cognitive engagement (CE) for which mathematics teachers employ digital tools can be investigated using the Interactive, Constructive, Active and Passive (ICAP) Theory. Our study aims to identify how different forms of intended CE influence each other and the role of age, gender, and teaching experience of mathematics teachers in relation to the ICAP constructs. In our study, we collected data from 687 mathematics teachers using the ICAP Technology Scale and conducted structural equation analyses. Our analyses indicate a high influence of Passive on Active CE (beta = .727) and Constructive on Interactive CE (beta = .964). Minor moderation effects were found regarding gender. Our study contributes to existing knowledge by providing insights into how different forms of intended CE among mathematics teachers interact in digitally supported classrooms. The results suggest that the professional development of mathematics teachers should emphasize flexible and hybrid approaches to integrating digital technologies and potentially enable direct transitions from lower to higher levels of CE.
We explored secondary school students' perceptions of digital educational tools in relation to basic psychological needs as outlined in Self-Determination Theory: autonomy, competence, and relatedness. Within a qualitative research design, we used a self-designed semi-standardized questionnaire to gather insights from a sample of 41 schoolchildren aged fifteen to seventeen from a school located in Aljustrel, in Baixo Alentejo, an economically modest region in southern Portugal. The findings reveal that students perceive the digital tools as effective in supporting their autonomy and competence, enabling self-paced, self-directed, and engaging learning experiences. However, students expressed mixed views on the fulfillment of relatedness, often preferring the interpersonal interactions and comprehensive explanations provided by in-person teaching. Based on the responses, digital tools may foster autonomy and competence, but their impact on relatedness is more context-dependent and less consistent. While digital tools can be effective for extracurricular interest development, students demonstrate a preference for teachers' personal attendance in formal school settings, especially when engrossed in a schedule of examinations.
