This study investigates how teacher- and school-level factors jointly shape teachers’ AI-TPACK (Artificial Intelligence - Technological Pedagogical Content Knowledge) under the digital transformation of education. Using machine learning and multilevel modeling to analyze data from 2166 teachers, we found AI-Ethics and AI-TCK were significant predictors of AI-TPACK. At the school level, diverse use of digital technologies moderated these associations. Additionally, schools promoting digital transformation tended to have teachers with stronger AI-TPACK. The findings highlight that the interaction between school digital strategies and teacher competence in technical knowledge contributes to AI-TPACK as an outcome of the ongoing digital transformation.

The development of speaking skills in English as a Foreign Language (EFL) is often hindered by limited practice opportunities, anxiety, and insufficient personalized feedback. Existing approaches typically employ single agent system (SAS) to handle tasks like grammar correction or pronunciation assessment. To address the fragmented functionality of SAS, we constructed a multi-agent system (MAS) for speaking practice. The system features a lightweight digital human interface and comprises 7 specialized agents responsible for preprocessing, dialogue generation, and dialogue supervision. A controlled experiment with university students demonstrated the superior efficacy of our MAS over SAS. Learners using MAS exhibited significantly greater gains in overall oral proficiency, with notable improvement in grammatical accuracy, a result attributed to the synergistic effects of hybrid input, contextual dialogue, and adaptive feedback. Although MAS received higher ratings on usability and ease of use, it did not significantly enhance perceived usefulness or willingness to communicate, partly due to response latency. Nonetheless, the findings confirm the strength of our MAS in delivering a pedagogically robust and emotionally supportive speaking practice environment. Future research will expand the validation scope and optimize agent coordination to better balance learning benefits with operational efficiency.

Abstract The rapid growth of massive online learning has intensified interest in self‐regulated learning (SRL) and socially shared regulation of learning (SSRL), yet empirical insights into their interplay in online collaborative learning (OCL) remain limited. This study employed a three‐layer analytical method combining cluster analysis, content analysis and complex network analysis to investigate how SSRL patterns evolve in relation to individual SRL profiles. Data from 60 undergraduates in a 16‐week course included over 16,000 trace entries (e.g. posts, replies, peer feedback) were collected and analysed. Results revealed that (1) three profiles of SRL were identified, based on learners' time investment, study regularity and help‐seeking behaviours; (2) groups with higher SSRL behavioural interaction displayed a more diverse and balanced role composition; and (3) distinct SSRL patterns emerged across SRL profiles over time. Individuals with high SRL profiles more frequently connected idea sharing with strategy using and process monitoring, while those with low SRL profiles relied on limited strategies focused mainly on idea sharing. These findings deepen the understanding of how individual regulation shapes group dynamics in online collaboration and suggest that instructional design should consider learners' SRL profiles when scaffolding their collaborative regulation processes. Practitioner Notes What is already known about this topic Self‐regulated learning (SRL) and socially shared regulation of learning (SSRL) are both essential for effective online collaborative learning (OCL), with prior research recognising their theoretical interplay. Trace data (e.g. login frequency, timestamp and posts) offer valuable opportunities to uncover how learners engage in and coordinate regulatory behaviours throughout online collaboration. What this paper adds Integrating SRL profiling with mixed network analysis, this study shows how individual SRL profiles drive divergent SSRL development, highlighting shared coordination patterns and evolving interactive roles across phases. Groups with higher SSRL interaction exhibited more diverse and balanced role composition, with central roles (e.g. leaders, animators) facilitating stronger collaborative regulation processes. Time‐series epistemic network analysis (ENA) provides a novel analytical lens to track the evolution of SSRL patterns across collaborative phases, linking SRL profiles with group‐level regulatory dynamics. Implications for practice and/or policy Tailoring support to SRL profiles allows educators to offer targeted scaffolds, using prompts and planning tools for low SRL learners, while encouraging high SRL learners to lead coordination and monitoring in SSRL. Role‐driven group design promotes collaboration structured by learners' SRL profiles.

Generative Artificial Intelligence (GenAI) is profoundly transforming the research field, with many researchers embracing its potential to enhance their work. However, the integration of GenAI into research practices still presents non-negligible challenges. To address these critical issues and better understand the current state of GenAI use in research, we conducted a systematic review focusing on GenAI-assisted research in the social sciences area and screened 8,831 relevant articles, of which 126 were retained for in-depth analysis. Our findings reveal that GenAI has been widely adopted in empirical studies across various social science domains, demonstrating its value in supporting research processes, such as conceptualization and research design, data collection and analysis, and writing and editing. Drawing on insights from existing practices, we developed a guideline for the academic community to support the development of a broader consensus on GenAI-assisted research practices, particularly in the field of education. This guideline aims to better regulate and leverage the potential of GenAI in research, ensuring its ethical and effective application.

Despite the rapid integration of Generative AI in K-12 education, significant disparities persist in teachers' actual usage behaviors. This study aims to predict teachers' AI usage behaviors and identify the primary determinants driving these behaviors. Utilizing large-scale data from OECD TALIS 2024, we applied machine learning algorithms to accurately predict teachers' AI usage behaviors. The results demonstrate that machine learning models outperform traditional linear models in predictive accuracy. Furthermore, SHAP interpretability analysis identifies AI attitude, AI training experience, and school innovative climate as the most critical predictors. These findings suggest that enhancing AI application requires not just technical provision, but critically hinges on targeted training and a supportive atmosphere.

Asynchronous online video learning has become an integral component of informal learning, while current research remains insufficient in examining the learning needs of asynchronous comments. Based on the long-tail theory, this study conducted a thematic evolution analysis on 22,273 asynchronous comments from humanities and science courses on a popular online platform. The findings revealed that the comments exhibit a typical long-tail tendency, where a small number of head comments form knowledge dissemination hubs, while a large volume of tail comments constitutes the foundation of micro-level interactions. Further analysis indicated that disciplinary differences profoundly shape engagement patterns, learners' comments on science courses demonstrate a clear technology-oriented tendency, whereas learners' comments on humanities courses rely more on emotional resonance. These findings provide learner engagement models based on disciplinary characteristics for the design of asynchronous learning videos, suggesting that platforms should establish differentiated interactive support systems.

As generative artificial intelligence (GenAI) rapidly evolves, K-12 education is introducing artificial intelligence (AI) courses whose interdisciplinary complexity exposes teachers' limited readiness. Research suggests that teachers can harness GenAI tools not only to enhance instruction but also to scaffold reflection that drives further teaching improvement. Therefore, this exploratory quasi-experimental study developed a customized GenAI agent system to support teachers' reflection on AI-course teaching and to enhance reflection self-efficacy, instructional design, and reflective thinking. A total of 60 in-service teachers were recruited and divided into two experimental groups, the self-reflection group (SRG) and the peer-reflection group (PRG), both supported by the customized GenAI agent system, and one control group (CG) with conventional technology-based reflection for a four-week AI course reflective practice experiment. The results revealed that the GenAI-agent-supported reflection approach could significantly promote teachers' self-reflection efficacy of two experimental groups compared with CG teachers. However, no significant differences between the SRG and PRG teachers' self-reflections efficacy could be found. Moreover, SRG and PRG teachers also outperformed the CG teachers on mutual-reflection self-efficacy. Furthermore, GenAI approach could significantly boost SRG and PRG teachers' instructional design reflection on methods and behavior; however, no significant differences were observed in instructional objectives or content among the three groups. To further explore the effects of GenAI agent support, epistemic network analysis was applied to examine the coded results of teachers' reflective journals. The findings indicated that SRG and PRG teachers demonstrated broader and higher order reflective thinking, integrating more dialogic and critical elements, whereas CG networks were predominantly descriptive. Overall, the study confirms that a customized GenAI agent can effectively deepen reflective thinking and practice, offering new insights into fostering teachers' professional development within K-12 AI education.

With the rapid development of information technology, online learning has become an integral part of modern education. Our study explores the long-tail distribution of learner behaviors in asynchronous learning video forums, focusing on platforms like Bilibili. By integrating Long Tail Theory with sentiment analysis, we propose a three-stage model of cognitive-emotional evolution, revealing content activation and social activation as dual driving factors influencing learner engagement. Our findings provide new insights for optimizing content recommendations and learner interactions in online education.

Cyberbullying has garnered growing attention, yet existing research lacks nuanced insights into the dynamics of students’ cyberbullying profiles and the associated risk factors across multiple domains. This study aims to (1) investigate K-12 students’ cyberbullying profiles, (2) develop an AI predictive model for cyberbullying roles, and (3) examine the relationship between cyberbullying profiles and risk factors across four domains: information and communication technology (ICT) profiles, moral development, normative social influence, and demographic characteristics. Latent profile analysis of 4721 students identified three distinct cyberbullying profiles: “Cyber bully-victims”, “Cyber passive and defenders”, and “Cyber victims and bystanders”. Through AI-based model competition experiments, we identified 13 key predictors and constructed a robust Random Forest model that accurately predicts profiles. The results reveal that information dissemination literacy emerges as the most prominent predictor, and students with higher literacy tend to fall into “Cyber bully-victims” or “Cyber victims and bystanders”. Additionally, students with excessively high moral emotions or difficulties in making moral judgments about harm are more prone to be “cyber victims and bystanders”. These findings reinforce a deeper understanding of cyberbullying profiles and their antecedents among K-12 students, and offer researchers an AI-based methodological approach for robust prediction of students’ cyberbullying in practice.

With the deepening of subject teaching practice supported by agents, human-machine collaborative learning has become the new normal of learning in the intelligent era. At present, there are insufficient researches on the influence mechanism of learnersu2019 psychological barriers in human-computer collaborative learning scenarios, which restricts the emotional design and optimization of educational agents. Therefore, this paper designed three dialogue scenarios: dialogue with subject agents, dialogue with general agents and dialogue with real person. Multimodal data encompassing electroencephalogram data, scale data, and interview data were collected during oral English communication experiments to systematically analyze the influence mechanism of different dialogue scenes on learnersu2019 psychological barriers in oral English communication. It was found out that there were obvious differences existed in the psychological barriers of oral communication between human-human and human-agent dialogue scenarios, and the effects of disciplinary agents and general agents in alleviating the psychological barriers of oral communication were also different. This research not only elucidated the unique advantages of human-agent collaborative dialogue model in mitigating psychological barriers in oral communication while providing empirical evidence for constructing an adaptive human-agent collaborative pedagogical framework based on discipline intelligent agents.

Knowledge tracing is a foundational task in intelligent education, aiming to predict students’ future performance by modeling their historical interactions. Traditional knowledge tracing methods primarily focus on analyzing students’ learning behaviors, often neglecting their dynamically changing emotional states during the learning process. Given that emotion is an integral part of the cognitive process, neglecting emotional states will limit the representation capacity of knowledge tracing models for modeling students’ learning processes. To address this issue, we propose a novel emotion-aware knowledge tracing method to explore the impact of complex emotions (e.g., concentration, frustration, boredom, and confusion) on knowledge states throughout the learning process. Specifically, we design an emotion-aware fusion module to capture the joint influence of multiple coexisting emotional states on students’ answers during each exercise session. Additionally, we utilize graph convolutional networks to propagate embeddings and obtain the complete knowledge structure information of exercises. We further integrate exercise attributes and students’ emotional information through a feature fusion module, thereby incorporating emotional states into the knowledge tracing task. Finally, we apply a transformer-based knowledge evolution module to model students’ evolving knowledge states, achieving multiview fusion modeling of emotions, knowledge topology, and learning sequences. Extensive experiments demonstrate that our method outperforms previous knowledge tracing methods in predicting student performance.

With the rapid development of artificial intelligence (AI) technology, AI-based multimodal learning analytics has attracted significant attention in the field of education. This paper proposes a framework for optimizing course design by integrating classroom multimodal data and large language models (LLMs). The framework combines classroom expression recognition, speech recognition, and LLM analysis to perform data fusion and intelligent analysis, enabling real-time course optimization. By integrating the RT-DETR and Whisper models for temporal modeling of multimodal data and combining ChatGPT with other large language models for course design improvement and evaluation, the optimal course design is selected. Experimental results show that this framework significantly enhances the accuracy of course design and the adaptability of teaching strategies, providing new insights for course design in dynamic educational environments.

Artificial Intelligence Learning Platform (AILP) plays a key role in AI education. However, there are few studies investigating the behavioral intentions of teachers and students on using AILP and the sample size is small. This study examines factors affecting the behavioral intentions of 299 teachers and 347 students from China on using AILP. "Perceived Playfulness" is integrated into the Unified Theory of Acceptance and Use of Technology (UTAUT) model as the theoretical framework. Moreover, this study analyzes the moderating effects of gender, age, experience, voluntariness of use, teachers' teaching level, teaching experience, and students' major. The results of research through two structural equation models show that: (1) Students focus more on performance expectancy, whereas teachers are more concerned with perceived playfulness. (2) Students are easily affected by social influence, while teachers are not. (3) Both teachers and students are impacted by effort expectancy and facilitating conditions. This research provides a scientific comparison of affecting factors about behavioral intentions on using AILP between teachers and students, which can be used by researchers and AILP designers to optimize AILP design for better AI education.

Although numerous studies have revealed various mechanisms involved in users' technology acceptance behaviors, the literature lacks insights about distinct profiles of teenagers' e-learning technology acceptance and how these profiles are associated with teenagers' contextual backgrounds and continued technology use. This study fills this gap by unpacking teenagers' learning management system (LMS) acceptance patterns through three-step latent profile analysis (LPA). The results of LPA on a survey of 1180 junior secondary students from 25 Hong Kong school identified three profiles, namely: Reluctant Pattern (8%), Embracing Pattern (61%), and Affirmative Pattern (31%), with ascending acceptance of LMS. A number of contextual factors were identified as antecedents of LMS acceptance, including school factors (i.e., school banding), family factors (i.e., family socioeconomic status (SES), family environment, and both parents' parenting styles), and personal factors (i.e., gender, experience, conceptions of learning, and ICT literacy). All the antecedents were significantly associated with profile membership, except for family SES and experience. In addition, the three acceptance patterns also affected consequences of LMS acceptance, represented by LMS continued use and use satisfaction. Students from Affirmative Pattern had higher satisfaction and more intensive LMS use than other acceptance patterns. The findings of this study not only contribute to theory through the development of teenagers' different latent profiles of LMS acceptance and relating these profiles to teenagers' contextual characteristics and LMS continued use, but also provide strong implications for future research on teenagers' e-learning technology use and for practitioners to improve LMS implementation and LMS use among teenagers.

学习投入度是影响教学的重要因素,其自动识别技术可以促进学生的自我调整和教师的教学改进.学习投入度自动识别的准确率主要受训练数据集的影响,在实践中常常因目标场景训练样本不足而难以训练出理想的识别模型.基于此,文章通过三组子实验,探讨了混合场景下的人脸数据集能否提高目标场景下学习投入度识别的准确率问题,结果发现:基于混合场景人脸数据集的学习投入度识别模型平均准确率高于其他两组基于非混合场景数据集的模型.此结论表明,在学习投入度识别任务中通过混合不同场景数据来扩大训练数据集,是解决当前目标场景下标注样本不足的有效策略.本研究对于开发精准的学习投入度识别模型有重要指导作用.

通过对教育信息化1.0时代的文献梳理与实践总结,指出教育信息化2.0时代高校智慧校园在实践中仍存在的典型问题,包括协同机制欠缺、教育数据治理体系不完善、师生个性化服务水平普遍低下、能充分彰显数据价值的智慧教育生态系统尚未形成.针对这些瓶颈问题,指出生态趋向是教育信息化2.0时代智慧校园建设的新动向,进而剖析了教育信息化2.0视域下智慧校园生态系统的建设策略,即应依托开放平台和微服务形成具有“生态”“人本”“智能”三个典型特征的新生态,从过往高耦合、低协同的1.0时代进化到低耦合、高协同的2.0新时代.

学习参与度是表征学生学习参与情况的重要指标,其自动识别方法是研究如何精确刻画学习效果变化动因及进行智慧教学决策的基础。现有研究已发现,学习参与度同情感投入、行为投入与认知投入存在直接关系,利用人工智能实现自动评估具有合理性与必要性。但相关研究数量有限,且主要聚焦在基于图像模态的表情识别领域。真实教学情境中的学习参与度识别应基于多模态数据的采集与分析,具体来说,可采用基于众包的方法建立多模态数据集,设计多模态融合的深度学习分析模型,并通过一致性检验完善模型的数据验证,以提高识别的准确率。因此,开展深度学习的实验研究具有较强的应用价值。该实验截选了中国大学MOOC网上的三门不同学科的视频片段,招募50位被试进行自主学习,每隔3秒自动记录被试脸部图像、脑电波数据和学习日志,初步建立了含3万余条面部表情图片和脑电波数据的多模态数据集,并基于后期融合策略及卷积神经网络结构中的ResNet架构,构建了一个多模态融合深度学习模型,进行模型训练。实验结果显示:该模型对未知被试的学习参与度预测的准确率可达87%,基于多模态的学习参与度识别方法,要优于基于单模态的学习参与度识别方法。

Previous research studies have demonstrated the influence of different user interface designs on student learning. This paper aims to examine the effect of different lecture video types on student learning using the eye-tracking and electroencephalography (EEG) techniques jointly. A two-factor experimental study was conducted on 62 undergraduate students who were randomly assigned to two groups, one of which used lecture video without teacher presence (Group 1, N = 31) and the other used lecture video with teacher presence (Group 2, N = 31). A survey was also conducted on both groups of students to collect data on their cognitive load and perceived satisfaction towards the lecture videos. The results of eye-tracking and EEG data indicate that teacher presence influences learners’ concentration and attention towards lecture video learning. Moreover, the learners’ perceived satisfaction are also related to students’ learning. Results of this study could provide future directions for studies on massively open online courses (MOOCs) and insights for MOOCs designers and educators to improve student learning with online courses.

计算思维是高中信息技术课程的四大学科核心素养之一.文章首先从分析计算思维的内涵出发,对计算思维进行了维度解析并构建了计算思维逻辑框架;在此基础上,文章提出了高中阶段计算思维的培养路径——研发培养计算思维的信息技术校本课程.随后,文章介绍了培养计算思维的高中信息技术校本课程的设计与开发过程.最后,文章阐述了校本课程的教学流程,分析了该流程中计算思维在解决问题环节的应用情况,并展示了相关的实践成果.文章的研究为计算思维在高中阶段的实施提供了一个系统化的完整案例,有助于深化培养学生计算思维的研究,并推动高中信息技术教学改革实践的深入开展.

As the intergenerational leap of education information, education informatization 2.0 should reconstruct the educa-tional ecosystem and make it have the typical characteristics of "humanitarianism", "ecology" and "intelligence". In the current re-search and practice of smart school, there is a tendency of conceptual generalization and blurred boundaries. Artificial intelligence is influencing and changing school ecosystem with clear path. From the construction of smart school to the construction of intelligence +school which has logic rationality and realistic demand. Therefore, we discuss the connotation and characteristics of "intelligence +"school from the perspective of "technology society". We also propose eight typical application scenarios of "intelligence +" school based on the current situation of AI application. "Intelligence +" school construction should adhere to the ecological strategy, open strategy and digital twin strategy based on data, so as to achieve all-round and systematic intelligence.