Frontiers in Education: Digital Learning Innovations
13 hours 7 minutes ago
BackgroundHandwriting is central to children's written-language development, but conventional assessment usually emphasizes the completed product and provides limited information about the processes that produced it. Digital technologies can capture handwriting as an unfolding motor and orthographic activity.ObjectiveThis scoping review mapped technologies for capturing children's handwriting process data, the raw signals and derived features obtained, the assessment constructs supported by process measures, and their educational, developmental, and clinical applications.MethodsFollowing scoping-review guidance and PRISMA-ScR, we searched Scopus, PubMed, ERIC, Web of Science Core Collection, and IEEE Xplore and conducted backward citation searching of 11 relevant review or framework articles. Of 198 reports sought for full-text assessment, one could not be retrieved, 197 were assessed, 39 were excluded, and 158 were included. All included reports were charted with the same 16-field framework. Independent verification by a second researcher covered randomly selected samples of 40 of 197 full-text eligibility decisions (20.3%; 97.5% agreement; Cohen's κ = 0.918) and 32 of 158 extractions (20.3%).ResultsDigitizing or graphics tablets were used in 136 reports (86.1%), and 117 (74.1%) used Wacom hardware. Position or trajectory (145, 91.8%) and time or timestamps (124, 78.5%) were the most common raw signals, followed by pen pressure or normal force (82, 51.9%) and pen-state or in-air information (71, 44.9%). Common derived feature families included speed or velocity (97, 61.4%), temporal or duration measures (84, 53.2%), spatial or geometric measures (67, 42.4%), and pressure- or force-derived measures (59, 37.3%). Developmental or normative characterization (75, 47.5%) and clinical characterization or differential assessment (74, 46.8%) were the most frequent applications. Latin-alphabetic handwriting appeared in 114 reports (72.2%), followed by Hebrew (23, 14.6%), Chinese (19, 12.0%), and Arabic (5, 3.2%).ConclusionsDigital handwriting assessment is well established in process capture but less consistently translated into applied educational or clinical decisions. Classification performance should not be equated with diagnostic validity. A theory-informed and review-refined four-level framework organizes raw signals, derived features, assessment constructs, and empirical interpretation or action without implying a validated developmental or psychometric hierarchy.
Siyuan Liu
13 hours 7 minutes ago
IntroductionDigital technologies and artificial intelligence (AI) are reshaping teacher education faster than many programmes can pedagogically respond. This scoping review examined how digital technologies are incorporated into teacher education, the advantages and barriers associated with these approaches, and how recent AI-focused research may help reinterpret persistent challenges.MethodsTwo analytically distinct Scopus-derived subsets were analysed. The RQ1/RQ2 subset comprised 155 studies on digitally supported teacher education, while the RQ3 subset included 40 AI-focused studies, yielding an overall evidence base of 195 publications. Data charting combined primary-category coding, descriptive frequency analysis, cross-category synthesis, and multi-label coding for the AI-focused evidence.ResultsWithin the RQ1/RQ2 subset, competence and ICT-integration frameworks were the most frequent strategy (46.5%). Digital competence development was the most frequently reported advantage (76.1%), whereas infrastructure and access represented the main barrier (45.2%). In the AI subset, professional development and teacher accompaniment (35.0%) and instructional design and assessment (32.5%) were the predominant approaches.DiscussionAcross both evidence streams, technological affordances became educationally meaningful when mediated by pedagogical design, disciplinary knowledge, critical and ethical judgement, human oversight, professional agency, and institutional support. The findings support an AI-responsive approach to teacher education that emphasises pedagogically justified and professionally accountable AI use rather than tool adoption alone.
John Fredy Patiño Hernández
13 hours 7 minutes ago
BackgroundDigital tools are increasingly integrated into higher education, yet evidence from Peruvian higher technological education remains limited.MethodsWe conducted an observational, analytical, cross-sectional study among 153 students from a public technological higher education institute in Callao, Peru. Data were collected in March–April 2025 using simple random sampling. Digital tools were assessed using an overall score and three domains: asynchronous interactive tools, synchronous interactive tools, and didactic resources. Meaningful learning was analyzed as a continuous score. Spearman correlations and linear regression models with HC3 robust standard errors were used. Multivariable models were adjusted a priori for sex and age group.ResultsMedian digital tools and meaningful learning scores were 94 (IQR: 85–103) and 85 (IQR: 75–96), respectively. Overall digital tools use correlated with meaningful learning (ρ=0.54; p < 0.001) and remained associated after adjustment (aβ=0.52; 95% CI: 0.41–0.64). Synchronous interactive tools showed the strongest independent association (aβ=0.81; 95% CI: 0.50–1.11), followed by didactic resources (aβ=0.46; 95% CI: 0.18–0.75). The association for asynchronous tools was attenuated after adjustment (aβ=0.21; 95% CI: −0.12 to 0.55; p = 0.213).ConclusionGreater digital tool use was associated with higher meaningful learning scores. Synchronous tools and didactic resources showed independent associations, whereas asynchronous tools did not retain an adjusted association. Longitudinal multicenter studies are needed to confirm these findings.
Edgar Pfuño-Ramos
2 days 12 hours ago
Digital entrepreneurship has emerged as one of the most transformative forces reshaping economies, labour markets, and educational expectations globally, yet initial teacher education (ITE) programmes have been conspicuously slow to respond. This article reports a systematic review of peer-reviewed literature published between 2015 and 2025, examining the extent to which digital entrepreneurship has been integrated into ITE curricula and what implications that integration, or its absence, carries for curriculum design. Searches were conducted across ERIC, Scopus, Web of Science, Google Scholar, and SciELO, yielding 312 initial results. Following rigorous inclusion and exclusion screening, 37 sources were retained and thematically analysed. Four interrelated themes emerged: the conceptual ambiguity surrounding digital entrepreneurship as an educational construct; the persistent gap between policy ambition and ITE curriculum reality; the inadequacy of pre-service teacher preparation for teaching digital entrepreneurship across conceptual, pedagogical, and technological dimensions; and the structural, cultural, and resource-related barriers that impede meaningful curriculum reform. Drawing on the EntreComp and DigCompEdu frameworks, the review concludes that ITE has not yet arrived at the destination its own policy documents have signposted. Fundamental redesign is required, not the incremental addition of technology modules, but a rethinking of what it means to prepare teachers for a digitally entrepreneurial world. The urgency of this task is compounded by the rise of generative artificial intelligence, acute youth unemployment across the Global South, and new evidence that pre-service teachers consistently underperform on precisely the digital competence dimensions most relevant to entrepreneurial practice. Five evidence-grounded curriculum implications are advanced: conceptual clarity before content delivery; integration rather than addition; the modelling imperative; assessment alignment; and structural partnership with the digital economy.
Beatrice Ngulube
5 days 12 hours ago
Generative artificial intelligence is increasingly used to produce formative feedback in programming education; however, whether students find such feedback useful, actionable, and worth using remains underexamined. This study presents a user-centred evaluation of feedback produced by an automated, rubric-based large language model assessment of programming responses, using a two-layer instrument, namely, response-level feedback on individual answers rated on five dimensions (clarity, specificity, accuracy, actionability, and usefulness) and a consolidated performance report rated on six dimensions (overall satisfaction, relevance, personalisation, cognitive load, intention to use, and motivation). Using a cross-sectional design, 144 students across secondary, short-cycle higher education (CTeSP—professional higher technical courses), and undergraduate programming rated 893 response-level feedback instances, and 140 of them rated 237 consolidated reports. Ratings were favourable across all dimensions, with student-level means of 4.24–4.43 (response level) and 4.11–4.38 (report level). At the response level, clarity and perceived accuracy were rated highest, and actionability and usefulness lowest. Cognitive load was the lowest at the report level. Only the between-context comparison for perceived accuracy reached statistical significance, with lower ratings in CTeSP (ε2 = 0.076); the small secondary sample limits conclusions about similarity. Perceived usefulness was most strongly associated with actionability and perceived accuracy (R2 = 0.76), and intention to use with motivation and personalisation (R2 = 0.53). These findings characterise the correlates of perceived usefulness and intention to use within each feedback layer, supporting human-centred designs that preserve student judgement and instructor oversight.
Pedro C. Mendonça
1 week ago
This study addresses the challenge of providing scalable and immersive environments for spoken English training in foreign language education. To this end, we designed and implemented a virtual reality (VR)-based instructional system that integrates immersive scenarios, avatar-based interaction, and task-based learning activities to support oral communication practice. A convergent parallel mixed-method design was employed to evaluate the effectiveness of the proposed instructional framework. Using convenience sampling, 60 university students from two intact English classes voluntarily participated in the study. One class (n = 30) received VR-supported instruction, whereas the other class (n = 30) received traditional instruction. Quantitative data were collected through pre- and post-intervention IELTS speaking assessments, while qualitative data were obtained from questionnaires and interviews to examine learners’ perceptions and experiences. Quantitative analyses, including independent-samples t-tests and ANCOVA, revealed no statistically significant difference in speaking proficiency between the two groups. However, a small effect size (Cohen's d = 0.2085) suggested a modest advantage for the VR group. Qualitative findings indicated high levels of learner engagement and perceived improvement, with over 90% of participants reporting increased confidence, vocabulary use, and opportunities for authentic communication. Learners particularly valued the immersive environments, avatar-based interaction, and immediate feedback provided through the VR platform. Rather than developing a new VR engine, this study contributed a structured instructional framework that demonstrates how existing immersive technologies can be systematically integrated into spoken English instruction. The findings provide practical insights into the pedagogical design, implementation, and classroom integration of VR technologies for language learning.
Hungche Chen
1 week 6 days ago
Generative artificial intelligence (GenAI) is rapidly transforming pedagogical practices in higher education by generating explanations, feedback, simulations, learning resources, and dialogic prompts. Existing AI frameworks in education predominantly conceptualize AI through functional roles, such as tutoring, assessment-centric models, institutional governance principles, or learner literacy perspectives. However, higher education institutions often regard GenAI as a complementary tool while simultaneously framing it as a threat to academic integrity, triggering reactive responses such as prohibition, surveillance, and detection. These framings leave a theoretical gap, offering limited insight into how GenAI redefines pedagogical agency, responsibility, and knowledge work in everyday interactions among instructors, students, and institutional structures. To address this gap, the present study proposes a nested instructor-student-GenAI triadic conceptual model for higher education. The model is derived through a focused integrative interdisciplinary synthesis that brings together literature from higher education, educational technology, learning sciences, instructional design, human-computer interaction, cognitive psychology, policy, ethics, and institutional governance. The model positions GenAI as a bounded didactic-pedagogical mediator operating within a shared didactic mediation space. Higher education institutions are conceptualized as the governance layer that enables, constrains, and legitimizes triadic practice through policies, infrastructure, regulations, and accountability mechanisms, while wider stakeholders shape external expectations. The study further formulates researchable propositions and discipline-sensitive implications to support future empirical validation and responsible GenAI integration in higher education.
Sharmila Rani Moganadas
1 week 6 days ago
IntroductionWith the growing integration of generative artificial intelligence (GenAI) into higher education, research attention must shift from the frequency of tool use to how these tools are used by learners. This study examines how GenAI literacy is related to learners' patterns of GenAI use and cognitive engagement during GenAI interactions.MethodsDrawing on questionnaire data from a large, diverse international sample of academic learners (N = 848), hierarchical multiple regression analyses were conducted to examine the relationships between dimensions of GenAI literacy and actual usage behaviors. Cluster analysis was then used to identify learner profiles based on patterns of GenAI use, literacy, monitoring, and reliance.ResultsThe findings indicate a dual role for critical-ethical awareness, a dimension of GenAI literacy that is associated with both monitoring GenAI outputs and using GenAI as a cognitive shortcut. Cluster analysis identified three learner profiles: Engaged and Literate users, who integrate monitoring and reliance; Uncritical Reliant users, who show high reliance but limited monitoring; and Skeptical Minimal users, who have low GenAI use but high reliance when they do engage. These patterns point to an “illusion of cognitive independence,” in which learners report high levels of independent thinking while engaging in cognitive offloading by relying on GenAI outputs without sufficient monitoring, particularly among Uncritical Reliant users.DiscussionCollectively, these findings suggest that GenAI literacy does not necessarily reduce reliance on GenAI but instead shapes how learners combine monitoring with reliance. Conceptually, this positions GenAI literacy as a regulatory mechanism of human–AI interaction, extending self-regulated learning to contexts in which cognitive processes are distributed between learners and systems. These results highlight the importance of developing GenAI literacy as a practice of reflective and regulated use rather than focusing solely on increasing technology adoption.
Meital Amzalag
2 weeks ago
Extended reality (XR) enables users to interact with and control simulated 3D elements, creating opportunities to support meaningful learning experiences, but it also introduces cognitive load due to its novel interactions. Researchers and designers have attempted to address these challenges by manipulating the user experience through the design of the virtual environment or by varying the level of interaction complexity. While traditional media benefit from established instructional frameworks, immersive media such as XR lack similarly refined design guidelines. In this perspective article, we argue that Mayer’s multimedia learning principles cannot be directly transferred to XR learning environments without reinterpretation, as XR introduces distinctive cognitive and interaction demands, including embodiment, spatial interaction, presence, agency, and environmental complexity. From this perspective, we organize XR design considerations around cognitive load processes and discuss how these choices may either support or hinder learning. We then propose key considerations for designing effective XR-based learning experiences, including content representation (e.g., 2D and 3D elements, color, positioning), virtual environment settings (e.g., background elements, haptic feedback, sound, voice interactions), and the structure of learning materials (e.g., pre-training materials, interactions and tasks, assessment).
Pedro Acevedo
2 weeks ago
This study determined the impact of a personalized educational platform based on machine learning (ML) on the cognitive learning of secondary education students at a public institution in Piura, Peru. The research followed an applied, quantitative approach with a quasi-experimental design. A sample of 56 fourth-grade students was distributed into a control group (n = 26, traditional teaching) and an experimental group (n = 30, platform-assisted teaching), using two pre-existing classrooms randomly assigned to each condition. Cognitive learning was assessed across four dimensions (attention, memory, reasoning, and comprehension) with pretest and posttest exams administered at the start and end of the same two-hour session per topic, over three consecutive weeks. The platform combined an unsupervised clustering model for cognitive profiling with a supervised classifier that predicted, in real time, the optimal content level for each student. Analyses were based on improvement scores (posttest - pretest); intergroup comparisons used the Mann–Whitney U test with rank-biserial effect sizes, and the relationship between platform metrics and the posttest score was examined through Spearman correlations. Results showed significant differences favoring the experimental group across the three topics (p < .001), with increasing effect sizes (r = 0.527, 0.610, 0.738), confirmed through a sensitivity analysis that controlled for the pretest. The experimental group achieved an average improvement of 8.29 points out of 20, versus 4.90 for the control group. Requested adaptations showed a moderate association on average with the posttest score (ρ = 0.485, p < .001), strengthening over the course of the intervention (up to ρ = 0.588), while the predictive weight of the initial micro-diagnostic declined progressively. The platform was associated with a significant, progressive, and dimension-differentiated improvement, although the single-classroom-per-condition design and the within-session measurement warrant interpretive caution. This work contributes to inclusive education in resource-limited contexts and aligns with Sustainable Development Goals 4 and 10.
Anel Enmita Córdova Zurita
2 weeks ago
IntroductionAlthough metacognitive scaffolding enhances self-regulated learning (SRL), its continuous impact on learners' internal cognitive states remains difficult to assess using traditional outcome-based measures.MethodsTo investigate these latent regulatory dynamics during a computer-supported learning task, we applied non-linear pupillary phase-space reconstruction to track the continuous cognitive engagement of university students (N = 82).ResultsLinear Mixed-Model analyses indicated a distinct divergence in pupillary dynamics: rather than uniformly reducing cognitive load, metacognitive prompts were associated with sustained physiological mobilization. Specifically, the scaffolded cohort demonstrated significantly elevated dynamic intensity (p < 0.001) and greater structural trajectory curvature (p < 0.001). Notably, these pupillary dynamics showed no significant linear association with final post-test scores (all p > 0.4).DiscussionWe interpret these pupillary dynamics as candidate physiological correlates of sustained regulatory micro-adjustment; however, because pupillometry is sensitive to arousal, effort, attention, and task-related factors, they cannot be equated with metacognitive monitoring directly. Immediate performance metrics alone may therefore fail to capture the physiological dynamics that accompany digital learning, although no definitive process–outcome dissociation can be inferred from the present data. By introducing non-linear pupillary dynamics as candidate continuous correlates of regulatory effort, this analytical framework provides a method for evaluating instructional designs and informing the development of real-time adaptive tutoring systems.
Kangning Xie
2 weeks 1 day ago
IntroductionGenerative artificial intelligence has become increasingly integrated into higher education; however, its adoption raises challenges related to authorship, transparency, content verification, privacy, and student responsibility. The objective of this study was to identify the factors associated with and structurally contributing to the ethical use of generative artificial intelligence among students at the State University of Milagro.MethodsA quantitative, applied, non-experimental, cross-sectional, correlational, explanatory-predictive, and confirmatory study was conducted. The sample comprised 980 students selected through non-probability purposive sampling. Data were collected using a 44-item questionnaire.ResultsThe proposed model demonstrated an excellent fit (CFI = 0.998; SRMR = 0.021; RMSEA = 0.008). The six predictor constructs jointly explained 44.0% of the variance in ethical GenAI use (R2 = 0.440), and all structural paths were positive and statistically significant. Academic integrity and transparency exhibited the strongest effect (β = 0.231), followed by ethical and technical literacy (β = 0.181), critical thinking and content verification (β = 0.177), institutional guidance and ethical education (β = 0.159), self-regulation and academic responsibility (β = 0.119), and data protection, privacy, and risk management (β = 0.114). Bivariate correlations between the six predictor constructs and the ethical use of generative artificial intelligence ranged from r = 0.370 to r = 0.507 (all p < 0.001).ConclusionThe ethical use of generative artificial intelligence was associated with the interaction of individual competencies, academic principles, and institutional conditions. These findings highlight the need to strengthen clear institutional policies, integrate ethics education across the curriculum, promote critical verification of AI-generated content, encourage transparent disclosure of AI use, and reinforce data protection practices. Given the cross-sectional design, the findings should be interpreted as predictive associations rather than evidence of causal relationships.
Yaima Beatriz Tabares-Cruz
2 weeks 6 days ago
BackgroundInclusive music-making requires instruments that support varied bodies, abilities, musical backgrounds, and forms of participation. Digital musical instruments provide diverse approaches to sound creation, and fabric-based interfaces offer an alternative interaction modality that may support participation for some users and contexts. Their tactile and deformable properties enable forms of interaction that differ from conventional rigid or screen-based controllers and may offer inclusive possibilities in particular settings.ObjectiveThis paper presents HarmonicThreads as a formative interaction-design case of a fabric-based digital musical instrument. The prototype explores how tactile cues, fabric deformation, projected visual feedback, and assisted accompaniment can support low-barrier music-making. We examine first-use discovery, perceived musical agency, and design considerations intended to inform the development of future fabric-based, accessible digital musical instruments.MethodsWe conducted a formative mixed-methods evaluation with 20 adult participants. Participants used HarmonicThreads in independent play, where conductive threads directly triggered notes, and assisted play, where sustained contact generated automated melodic accompaniment. Data included observations, SUS and UEQ-S responses, semi-structured interviews, and participant-generated design sketches.ResultsParticipants treated the fabric surface as an instrument, using raised threads to identify playable regions, form spatial pitch expectations, and experiment with chords, rhythm, and sustained notes. Qualitative results show that assisted play supported immediate musical engagement, while independent play supported greater perceived control. The study also revealed accessibility-relevant barriers, including uncertain activation thresholds, perceived fragility, thin and closely spaced threads, posture and fatigue, ambiguous visual feedback, and limited expressive mappings.ConclusionHarmonicThreads contributes a formative design case for inclusive digital music technology. We identify design considerations for tactile legibility, low-effort activation, ergonomic support, transparent multimodal feedback, configurable scaffolding, richer sonic mappings, and future participatory evaluation with intended users and real music-making contexts.
Ellie Nguyen
3 weeks 5 days ago
This study examines how interactive digital mnemonics reshape the temporal dynamics of associative retrieval during foreign-language vocabulary acquisition among adolescent learners. A quasi-experimental design compared an adaptive digital learning environment with conventional paper-based methods in a sample of 102 middle-school students (aged 12–14), assigned at the intact-class level to an experimental group (n = 51) and a control group (n = 51). Participants studied 45 health- and sport-related English lexical items. A computerized lexical decision task programmed in PsychoPy measured reaction time (RT) in milliseconds and response accuracy across three phases: baseline pre-test (T0), immediate post-test (T1), and four-week delayed post-test (T2). The experimental group exhibited a statistically significant reduction in RT latencies at T1 [M = 942 ms vs. 1,520 ms; F(2, 200) = 84.30, p < 0.001, η²p = 0.46] and maintained faster retrieval at T2 (1,055 ms vs. 1,710 ms). Accuracy at T2 remained substantially higher in the experimental cohort (81.5% vs. 54.1%), indicating greater retention resilience. These findings provide behavioral evidence that the integrated digital mnemonic package was associated with faster associative retrieval and more durable L2 vocabulary retention. Because cognitive load was not measured directly, the results are interpreted as evidence of improved retrieval efficiency rather than as direct confirmation of reduced extraneous load.
Zhanar Ibrayeva
3 weeks 6 days ago
Teaching students to solve geometric problems is one of the most pressing issues in mathematics education. Problem-solving is a primary means of applying students’ theoretical knowledge and constitutes an essential part of their work in geometry lessons. However, classroom practice indicates that many students experience serious difficulties in solving mathematical problems and often fail to develop adequate problem-solving skills. These difficulties are particularly evident in geometry. Therefore, further investigation is needed into instructional approaches that may improve students’ geometric problem-solving abilities. This study aimed to develop a GeoGebra-supported methodology for building students’ foundational knowledge and to test its effectiveness in teaching geometric problem-solving. A total of 370 students from four schools in Almaty—Schools Nos. 17, 33, 148, and 172—participated in the study. The experimental and control groups each consisted of 185 students from classes with similar levels of prior geometry achievement, based on their results from the previous academic year. The findings showed that students in the experimental group demonstrated higher levels of foundational knowledge and better problem-solving performance than those in the control group. The GeoGebra-supported methodology also contributed to students’ understanding of geometric concepts and improved their ability to solve geometric problems.
Almagul Ardabayeva
3 weeks 6 days ago
Sergio Ruiz-Viruel
4 weeks ago
This study investigates the effectiveness of integrating Artificial Intelligence (AI) within a Project-Based Inquiry (PBI) framework to enhance elementary students’ mathematical connections in geometry learning. AI functions as a cognitive scaffolding tool by providing adaptive prompts that support students’ reasoning, representation, and reflection, while PBI engages learners in authentic problem-solving tasks such as designing classroom layouts and park environments.The indicators of mathematical connections assessed in this study include: (1) connections between mathematical concepts (e.g., area and multiplication), (2) connections across multiple representations (visual, symbolic, and verbal), and (3) connections between mathematics and real-world contexts. This study employed a Design-Based Research (DBR) approach involving iterative cycles of design, implementation, evaluation, and refinement. Data were collected at the elementary school level through classroom observations, students’ project artifacts, short interviews, and AI–student interaction logs. The findings reveal substantial improvements in students’ mathematical communication and reasoning skills. Quantitative results indicate increases across all assessed dimensions, including precision of vocabulary usage (from 38% to 76%), logical structuring of arguments (from 42% to 71%), interactive discourse (from 36% to 68%), and accuracy in problem-solving (from 44% to 79).Qualitative findings further show that students transitioned from procedural to conceptual reasoning, demonstrating the ability to connect geometric concepts with real-world design contexts. AI-supported scaffolding facilitated this shift by guiding students from informal reasoning toward formal mathematical modeling. Interview data confirmed increased student engagement, metacognitive awareness, and confidence in problem-solving.These results suggest that AI-supported PBI effectively promotes interconnected mathematical understanding and meaningful learning experiences in elementary education.
Fery Muhamad Firdaus
4 weeks ago
Against the backdrop of the ongoing technological revolution and industrial transformation, artificial intelligence (AI) has emerged as a pivotal technological driver for advancing the high-quality development of higher education. As one of China's leading provinces in both higher education and manufacturing, Shandong faces a pressing need to integrate AI into its higher education system to facilitate educational modernization and support the broader transition from old to new growth drivers. This study examines the theoretical underpinnings and operational mechanisms through which AI contributes to the high-quality development of higher education and explores its practical implications within the specific context of Shandong Province. Focusing on key dimensions—including talent cultivation, teaching and learning innovation, university governance, and the optimization of educational resource allocation—the study demonstrates that AI enhances higher education development primarily through data-driven decision-making, intelligent technologies, and human–machine collaboration. However, the empirical application of AI within Shandong's higher education sector remains constrained by fragmented technological implementation, insufficient AI literacy among faculty members, and underdeveloped institutional and ethical governance frameworks. To address these challenges, this paper proposes a series of policy recommendations, including strengthening top-level institutional design, promoting the intelligent transformation of teaching and learning processes, enhancing faculty capacity building, and establishing robust systems for AI governance and ethical regulation. The findings offer both theoretical insights and policy implications for advancing AI-enabled high-quality development of higher education at the regional level.
Cuihua An
4 weeks ago
This study investigates the application of generative artificial intelligence (generative AI) in mathematics education from 2022 to 2026. Drawing on a systematic literature review of 268 articles retrieved from OpenAlex, and EBSCO databases, we identify and analyze 66 core research articles through a rigorous screening process following PRISMA guidelines. The findings reveal three key patterns. First, generative AI applications in mathematics education span four domains: student learning support (64%), teacher instruction assistance (38%), assessment and feedback innovation (27%), and special education (2%), with student learning support as the dominant application area. Second, while generative AI shows positive effects in enhancing learning motivation and providing personalized support, it exhibits notable limitations in mathematical reasoning accuracy and critical thinking development—with 70% of studies documenting AI limitations. Third, challenges associated with these applications are multifaceted, encompassing technical (30%), pedagogical (52%), ethical (20%), and implementation (42%) dimensions. We propose an integrated analytical framework combining the Technological Pedagogical Content Knowledge (TPACK) framework with hybrid intelligence theory, and identify a novel pedagogical paradigm in which AI errors serve as instructional resources for cultivating critical thinking. Future research directions include broadening application domains, deepening theoretical frameworks, and strengthening empirical investigation. This review is limited to English-language publications from two databases, with the search closing in June 2026. Despite these limitations, the review contributes systematic evidence and practical guidance for the effective integration of generative AI in mathematics education.
Zengfu Chao
4 weeks ago
This study examines lecturer-guided artificial intelligence (AI)-supported flipped learning in undergraduate Information Technology programming education. The study focuses on how multiple AI tools can be pedagogically orchestrated across pre-class preparation, in-class programming practice, and post-class reflection while maintaining lecturer guidance, student reasoning, and responsible AI use. A classroom-based non-equivalent groups quasi-experimental design was implemented with 120 undergraduate Information Technology students in a Programming Techniques course. One intact class participated in a lecturer-guided AI-supported flipped learning condition (n = 60), while another intact class participated in a traditional flipped classroom condition without planned AI-supported learning tasks (n = 60). Quantitative data included pretest and posttest programming scores and post-intervention self-reported measures of learning autonomy, reflective learning, and digital competence. Qualitative interview data from students and lecturers were used to contextualize participants' perceptions of AI-supported learning activities. The results showed that the AI-supported flipped learning group achieved higher posttest programming performance, showed a larger observed programming-performance gain, and reported higher post-intervention learning autonomy, reflective learning, and digital competence than the traditional flipped classroom group. Because the self-reported items referred explicitly to AI-supported activities, these differences are interpreted as condition-linked perceptions rather than as comparisons of general competence. Qualitative findings indicated that students and lecturers perceived AI as useful for preparation, debugging-related feedback, solution comparison, reflection, and lecturer-guided verification. The findings should be interpreted as classroom-based evidence of observed group differences rather than as definitive causal proof of AI tool effects. The main contribution of the study is to illustrate how multiple AI tools can be organized as a lecturer-guided pedagogical orchestration within the flipped programming learning cycle.
Ngoc Duong Bach