As programming education scales, evaluating student code becomes increasingly challenging. In object-oriented programming (OOP) courses, design patterns are crucial for teaching maintainable reusable solutions that meet industry standards. While traditional automated assessment tools have successfully addressed code correctness and quality, automating the detection of design patterns presents a distinct challenge. Current methods, such as static code analysis combined with graph similarity, have proven effective for library code but often struggle with the variability of student submissions. This article investigates the application of generative artificial intelligence (GenAI) models to improve accuracy in detecting design patterns in student-written code. Our research addresses two key questions: 1) Which of the current GenAI models offer optimal performance, accuracy, and reasoning capabilities for design pattern assessment? and 2) How does a cloud-based Software as a Service solution (such as ChatGPT) compare to a cloudlet solution (local model deployed on the University’s cluster) in terms of reliability and scalability? We assess the effectiveness of these approaches using a representative sample of student assignments specifically crafted to require design pattern implementation. Our findings discuss the educational utility of GenAI in reducing instructors’ grading burdens, enhancing students’ self-assessment opportunities, and its potential to guide industry practitioners in design pattern evaluation. We highlight human-in-the-loop supervision to ensure that GenAI recommendations align with educational standards and advocate for open-source cloudlets as a scalable cost-effective infrastructure for universities. Our study provides guidelines for integrating GenAI into OOP coursework and highlights the implications for teaching, industry applicability, and future research directions.
{"title":"Using GenAI to Assess Design Patterns in Student-Written Code","authors":"Daniel-Florin Dosaru;Diana-Maria Simion;Andrei-Horia Ignat;Lorina-Cristina Negreanu;Alexandru-Corneliu Olteanu","doi":"10.1109/TLT.2025.3604054","DOIUrl":"https://doi.org/10.1109/TLT.2025.3604054","url":null,"abstract":"As programming education scales, evaluating student code becomes increasingly challenging. In object-oriented programming (OOP) courses, design patterns are crucial for teaching maintainable reusable solutions that meet industry standards. While traditional automated assessment tools have successfully addressed code correctness and quality, automating the detection of design patterns presents a distinct challenge. Current methods, such as static code analysis combined with graph similarity, have proven effective for library code but often struggle with the variability of student submissions. This article investigates the application of generative artificial intelligence (GenAI) models to improve accuracy in detecting design patterns in student-written code. Our research addresses two key questions: 1) Which of the current GenAI models offer optimal performance, accuracy, and reasoning capabilities for design pattern assessment? and 2) How does a cloud-based Software as a Service solution (such as ChatGPT) compare to a cloudlet solution (local model deployed on the University’s cluster) in terms of reliability and scalability? We assess the effectiveness of these approaches using a representative sample of student assignments specifically crafted to require design pattern implementation. Our findings discuss the educational utility of GenAI in reducing instructors’ grading burdens, enhancing students’ self-assessment opportunities, and its potential to guide industry practitioners in design pattern evaluation. We highlight human-in-the-loop supervision to ensure that GenAI recommendations align with educational standards and advocate for open-source cloudlets as a scalable cost-effective infrastructure for universities. Our study provides guidelines for integrating GenAI into OOP coursework and highlights the implications for teaching, industry applicability, and future research directions.","PeriodicalId":49191,"journal":{"name":"IEEE Transactions on Learning Technologies","volume":"18 ","pages":"869-876"},"PeriodicalIF":4.9,"publicationDate":"2025-08-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11144915","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145351916","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-08-29DOI: 10.1109/TLT.2025.3604096
Nguyen Nang Hung Van;Phuc Hao Do;Van Nam Hoang;Truc Thi Kim Nguyen;Minh Tuan Pham
This study investigates how technical advances in large language models (LLMs) translate into measurable educational benefit. University admission counseling plays a crucial role in helping prospective students make their higher education decisions. However, traditional advisory methods are constrained by issues, such as limited scalability, personalization, and the ability to handle large volumes of inquiries. With the growing need for real-time assistance, artificial intelligence (AI), particularly LLMs), presents a promising solution to these challenges. This article introduces an AI-driven university admission counseling system that automates routine inquiries, personalizes guidance, and improves accessibility. We develop a formal mathematical framework to represent the counseling task, using embedded and similarity metrics to assess the compatibility of student profiles with academic programs. The system incorporates a multistage workflow for efficient data processing, embedded generation, and AI-driven recommendation. We evaluated the performance of several LLMs, namely, eLLAMA, eGPT, and eDEEPSEEK, through retrieval-augmented generation, measuring output quality with natural language processing metrics, such as bilingual evaluation understudy, recall-oriented understudy for gisting evaluation, METEOR, and BERTScore. Our results demonstrate that LLMs can significantly improve the efficiency and quality of admission counseling, providing a scalable and adaptable solution that demonstrably enhances student confidence and decision quality.
{"title":"AI-Powered University Admission Counseling: A Use Case of Large Language Models in Student Guidance","authors":"Nguyen Nang Hung Van;Phuc Hao Do;Van Nam Hoang;Truc Thi Kim Nguyen;Minh Tuan Pham","doi":"10.1109/TLT.2025.3604096","DOIUrl":"https://doi.org/10.1109/TLT.2025.3604096","url":null,"abstract":"This study investigates how technical advances in large language models (LLMs) translate into measurable educational benefit. University admission counseling plays a crucial role in helping prospective students make their higher education decisions. However, traditional advisory methods are constrained by issues, such as limited scalability, personalization, and the ability to handle large volumes of inquiries. With the growing need for real-time assistance, artificial intelligence (AI), particularly LLMs), presents a promising solution to these challenges. This article introduces an AI-driven university admission counseling system that automates routine inquiries, personalizes guidance, and improves accessibility. We develop a formal mathematical framework to represent the counseling task, using embedded and similarity metrics to assess the compatibility of student profiles with academic programs. The system incorporates a multistage workflow for efficient data processing, embedded generation, and AI-driven recommendation. We evaluated the performance of several LLMs, namely, eLLAMA, eGPT, and eDEEPSEEK, through retrieval-augmented generation, measuring output quality with natural language processing metrics, such as bilingual evaluation understudy, recall-oriented understudy for gisting evaluation, METEOR, and BERTScore. Our results demonstrate that LLMs can significantly improve the efficiency and quality of admission counseling, providing a scalable and adaptable solution that demonstrably enhances student confidence and decision quality.","PeriodicalId":49191,"journal":{"name":"IEEE Transactions on Learning Technologies","volume":"18 ","pages":"856-868"},"PeriodicalIF":4.9,"publicationDate":"2025-08-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145141689","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-08-19DOI: 10.1109/TLT.2025.3600112
Ye Zhang;Mo Wang;Jinlong He;Yupeng Zhou;Hongping Wu;Zhaoyang Sun;Yujie Zhang;Minghao Yin
Aesthetic perception, as a core competence in art education, fosters students’ cultural sensibility, emotional expression, and critical thinking. However, existing approaches to cultivating aesthetic perception often lack systematic guidance and personalized developmental pathways, limiting their capacity to support sustained and individualized growth. Two central challenges remain unresolved: first, how to effectively model the dynamic, multidimensional progression of students’ aesthetic understanding, and second, how to construct coherent learning paths that guide students from basic perceptual awareness to more abstract artistic engagement. To address these issues, we propose AesthPath a reinforcement learning-based recommendation model that constructs personalized picture book learning paths to enhance aesthetic perception. Specifically, the model introduces a Markov decision process formulation that captures the evolving states of learners’ aesthetic competence across multiple dimensions. An actor–critic algorithm is then employed to generate adaptive learning trajectories by balancing exploration of new content with the reinforcement of effective materials, based on ongoing learner feedback. Unlike traditional static or rule-based recommendation methods, AesthPath supports fine-grained, feedback-driven optimization of learning trajectories, facilitating goal-oriented and personalized development of aesthetic perception. Experimental results on a real-world dataset demonstrate the effectiveness of AesthPath in enhancing students’ aesthetic understanding. This study offers new theoretical and methodological insights for intelligent learning path design and educational recommendations, highlighting the potential of reinforcement learning in adaptive learning scenarios.
{"title":"Reinforcement Learning-Driven Optimization of Picture Book Paths for Aesthetic Perception Enhancement","authors":"Ye Zhang;Mo Wang;Jinlong He;Yupeng Zhou;Hongping Wu;Zhaoyang Sun;Yujie Zhang;Minghao Yin","doi":"10.1109/TLT.2025.3600112","DOIUrl":"https://doi.org/10.1109/TLT.2025.3600112","url":null,"abstract":"Aesthetic perception, as a core competence in art education, fosters students’ cultural sensibility, emotional expression, and critical thinking. However, existing approaches to cultivating aesthetic perception often lack systematic guidance and personalized developmental pathways, limiting their capacity to support sustained and individualized growth. Two central challenges remain unresolved: first, how to effectively model the dynamic, multidimensional progression of students’ aesthetic understanding, and second, how to construct coherent learning paths that guide students from basic perceptual awareness to more abstract artistic engagement. To address these issues, we propose AesthPath a reinforcement learning-based recommendation model that constructs personalized picture book learning paths to enhance aesthetic perception. Specifically, the model introduces a Markov decision process formulation that captures the evolving states of learners’ aesthetic competence across multiple dimensions. An actor–critic algorithm is then employed to generate adaptive learning trajectories by balancing exploration of new content with the reinforcement of effective materials, based on ongoing learner feedback. Unlike traditional static or rule-based recommendation methods, AesthPath supports fine-grained, feedback-driven optimization of learning trajectories, facilitating goal-oriented and personalized development of aesthetic perception. Experimental results on a real-world dataset demonstrate the effectiveness of AesthPath in enhancing students’ aesthetic understanding. This study offers new theoretical and methodological insights for intelligent learning path design and educational recommendations, highlighting the potential of reinforcement learning in adaptive learning scenarios.","PeriodicalId":49191,"journal":{"name":"IEEE Transactions on Learning Technologies","volume":"18 ","pages":"798-811"},"PeriodicalIF":4.9,"publicationDate":"2025-08-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145021155","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The initial version of ProgTutor, a learning framework designed to teach the fundamentals of computer programming in a personalized and applied manner, is presented here. The main contribution of this tool is the integration of an adaptive learning system with a 3-D robotic simulator, used to face realistic challenges in a user-friendly fashion. ProgTutor provides automated evaluations and feedback on coding errors, ensuring that learners receive the support they need to progress effectively. In addition, it features dynamic learning paths tailored to each student’s pace, offloading tasks such as automated evaluation and adaptive sequencing to the tool so that students and teachers can concentrate on judgment. ProgTutor also enhances the teachers’ capacities as educators, as they can focus their attention on those students with more learning difficulties. Therefore, it functions as intelligence augmentation rather than automation, with teachers remaining in the decision loop. This article introduces the conceptual and functional design of ProgTutor, and it includes piloting results with high school students during the academic course 2023–2024, focused on their acceptability of the tool and on the analysis of the real impact that this type of system could have over the formal educational landscape in the future.
{"title":"ProgTutor: A Robotic-Based Framework to Support Teaching and Learning of Programming Fundamentals","authors":"Javier Ortega-Morla;Antonio Leis;Alma Mallo;Laura Morán-Fernández;Sara Guerreiro;Alejandro Paz-López;Beatriz Pérez-Sánchez;Noelia Sánchez-Maroño;Alejandro Rodríguez-Arias;Oscar Fontenla-Romero;Francisco Bellas","doi":"10.1109/TLT.2025.3598041","DOIUrl":"https://doi.org/10.1109/TLT.2025.3598041","url":null,"abstract":"The initial version of ProgTutor, a learning framework designed to teach the fundamentals of computer programming in a personalized and applied manner, is presented here. The main contribution of this tool is the integration of an adaptive learning system with a 3-D robotic simulator, used to face realistic challenges in a user-friendly fashion. ProgTutor provides automated evaluations and feedback on coding errors, ensuring that learners receive the support they need to progress effectively. In addition, it features dynamic learning paths tailored to each student’s pace, offloading tasks such as automated evaluation and adaptive sequencing to the tool so that students and teachers can concentrate on judgment. ProgTutor also enhances the teachers’ capacities as educators, as they can focus their attention on those students with more learning difficulties. Therefore, it functions as intelligence augmentation rather than automation, with teachers remaining in the decision loop. This article introduces the conceptual and functional design of ProgTutor, and it includes piloting results with high school students during the academic course 2023–2024, focused on their acceptability of the tool and on the analysis of the real impact that this type of system could have over the formal educational landscape in the future.","PeriodicalId":49191,"journal":{"name":"IEEE Transactions on Learning Technologies","volume":"18 ","pages":"783-797"},"PeriodicalIF":4.9,"publicationDate":"2025-08-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11122894","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144909334","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-08-05DOI: 10.1109/TLT.2025.3596125
Xi Bei Xiong;Simin Cao;Tianhang Gao;Xiya Feng;Hui Li
This study examined early childhood teachers’ artificial intelligence (AI) literacy in Guangxi, China. Utilizing data from 1522 kindergarten teachers, we developed and validated a culturally adapted AI literacy scale through factor analyses, confirming a three-construct structure: Safety, Attitude, and Capability. Latent profile analysis identified three distinct teacher profiles: “Hesitant Beginners” (9.6%), “Enthusiastic Practitioners” (64.2%), and “Confident Experts” (26.2%), revealing significant heterogeneity. Teachers generally exhibited positive attitudes toward AI but lower safety awareness and capability levels. Regression analyses indicated that education level, working experience (negatively associated), kindergarten type, and geographic location (urban/rural) significantly influence AI literacy levels and profile membership. These findings underscore the need for context-specific assessment tools and tailored teacher education programs to enhance their digital literacy and promote equitable AI integration in Chinese early childhood education.
{"title":"From Hesitant Beginners to Confident Experts: Profiles and Predictors of AI Literacy Among Preschool Teachers in Guangxi, China","authors":"Xi Bei Xiong;Simin Cao;Tianhang Gao;Xiya Feng;Hui Li","doi":"10.1109/TLT.2025.3596125","DOIUrl":"https://doi.org/10.1109/TLT.2025.3596125","url":null,"abstract":"This study examined early childhood teachers’ artificial intelligence (AI) literacy in Guangxi, China. Utilizing data from 1522 kindergarten teachers, we developed and validated a culturally adapted AI literacy scale through factor analyses, confirming a three-construct structure: Safety, Attitude, and Capability. Latent profile analysis identified three distinct teacher profiles: “Hesitant Beginners” (9.6%), “Enthusiastic Practitioners” (64.2%), and “Confident Experts” (26.2%), revealing significant heterogeneity. Teachers generally exhibited positive attitudes toward AI but lower safety awareness and capability levels. Regression analyses indicated that education level, working experience (negatively associated), kindergarten type, and geographic location (urban/rural) significantly influence AI literacy levels and profile membership. These findings underscore the need for context-specific assessment tools and tailored teacher education programs to enhance their digital literacy and promote equitable AI integration in Chinese early childhood education.","PeriodicalId":49191,"journal":{"name":"IEEE Transactions on Learning Technologies","volume":"18 ","pages":"812-821"},"PeriodicalIF":4.9,"publicationDate":"2025-08-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145078678","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-07-24DOI: 10.1109/TLT.2025.3592579
Jarrett E. Woo;Jin Woo Kim;Kwangtaek Kim;Jeremy M. Jarzembak;Ann James;Jennifer Biggs;John Dunlosky;Robert Clements
Training healthcare professionals in intravenous (IV) needle insertion is a critical component of medical education, traditionally relying on manikin-based simulations and real-life practice. However, the advent of haptic virtual reality (HVR) technologies offers a transformative approach to this training, enhancing safety and potential efficiency. This study explores the development of an IV needle insertion simulator using two different haptic devices integrated into a VR system on the Unity platform and assesses its impact on learning through a three-week experiment. The simulator is designed to create a realistic and immersive training environment by incorporating detailed anatomical models, physics-based hand interactions, and real-time haptic feedback. The virtual environment replicates a clinical setting, featuring a patient arm model and an IV needle. The haptic feedback is programmed to offer realistic feelings of needle insertion and hand grasping, improving the user’s accuracy. Learning impact and usability testing with 41 students indicate a promising improvement in skill acquisition and confidence. Specifically, participants showed a 55% increase in success rates and a significant boost in confidence. This high-fidelity HVR simulator represents a significant step forward in medical training technologies, offering a scalable and repeatable training tool adaptable to various educational needs and skill levels.
{"title":"Enhancing IV Needle Insertion Training With a Bimanual Haptic VR Simulator: Development, Usability, and Learning Impact","authors":"Jarrett E. Woo;Jin Woo Kim;Kwangtaek Kim;Jeremy M. Jarzembak;Ann James;Jennifer Biggs;John Dunlosky;Robert Clements","doi":"10.1109/TLT.2025.3592579","DOIUrl":"https://doi.org/10.1109/TLT.2025.3592579","url":null,"abstract":"Training healthcare professionals in intravenous (IV) needle insertion is a critical component of medical education, traditionally relying on manikin-based simulations and real-life practice. However, the advent of haptic virtual reality (HVR) technologies offers a transformative approach to this training, enhancing safety and potential efficiency. This study explores the development of an IV needle insertion simulator using two different haptic devices integrated into a VR system on the Unity platform and assesses its impact on learning through a three-week experiment. The simulator is designed to create a realistic and immersive training environment by incorporating detailed anatomical models, physics-based hand interactions, and real-time haptic feedback. The virtual environment replicates a clinical setting, featuring a patient arm model and an IV needle. The haptic feedback is programmed to offer realistic feelings of needle insertion and hand grasping, improving the user’s accuracy. Learning impact and usability testing with 41 students indicate a promising improvement in skill acquisition and confidence. Specifically, participants showed a 55% increase in success rates and a significant boost in confidence. This high-fidelity HVR simulator represents a significant step forward in medical training technologies, offering a scalable and repeatable training tool adaptable to various educational needs and skill levels.","PeriodicalId":49191,"journal":{"name":"IEEE Transactions on Learning Technologies","volume":"18 ","pages":"769-782"},"PeriodicalIF":4.9,"publicationDate":"2025-07-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144868433","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In e-learning, high-quality learning path generation can meet learners’ personalized demands and solve their cognitive disorientation dilemma. However, existing learning path generation schemes still have challenges, such as focusing solely on one aspect of the learner’s characteristics or the structure of learning content, difficulty in describing the variation in a learner’s knowledge level, and a lack of interpretability. To address these issues, in this article, we propose an educational psychology-empowered personalized learning path generation strategy. First, inspired by Brown’s decay theory of immediate memory, we design the decay attentive knowledge tracing approach for assessing a learner’s knowledge level. Then, motivated by Bruner’s cognitive structure learning theory, we present search space optimization for selecting the learning content candidate set. Finally, enlightened by Posner’s conceptual change model, we impose multiple rule constraints on the matching process of the learner’s knowledge level and the learning content in the candidate set, gradually forming the personalized learning path. Experimental results demonstrate the performance of the proposed strategy for guaranteeing the reasonableness of learning content organization and enhancing the learner’s knowledge level. Moreover, the actual utilization of the proposed strategy in higher education course instruction shows its effectiveness in improving learning outcomes, motivation, and engagement.
{"title":"Educational Psychology-Empowered Personalized Learning Path Generation Strategy","authors":"Xin Wei;Wenrui Han;Shiyun Sun;Junhao Shan;Liang Zhou","doi":"10.1109/TLT.2025.3590602","DOIUrl":"https://doi.org/10.1109/TLT.2025.3590602","url":null,"abstract":"In e-learning, high-quality learning path generation can meet learners’ personalized demands and solve their cognitive disorientation dilemma. However, existing learning path generation schemes still have challenges, such as focusing solely on one aspect of the learner’s characteristics or the structure of learning content, difficulty in describing the variation in a learner’s knowledge level, and a lack of interpretability. To address these issues, in this article, we propose an educational psychology-empowered personalized learning path generation strategy. First, inspired by Brown’s decay theory of immediate memory, we design the decay attentive knowledge tracing approach for assessing a learner’s knowledge level. Then, motivated by Bruner’s cognitive structure learning theory, we present search space optimization for selecting the learning content candidate set. Finally, enlightened by Posner’s conceptual change model, we impose multiple rule constraints on the matching process of the learner’s knowledge level and the learning content in the candidate set, gradually forming the personalized learning path. Experimental results demonstrate the performance of the proposed strategy for guaranteeing the reasonableness of learning content organization and enhancing the learner’s knowledge level. Moreover, the actual utilization of the proposed strategy in higher education course instruction shows its effectiveness in improving learning outcomes, motivation, and engagement.","PeriodicalId":49191,"journal":{"name":"IEEE Transactions on Learning Technologies","volume":"18 ","pages":"741-756"},"PeriodicalIF":4.9,"publicationDate":"2025-07-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144739961","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-07-10DOI: 10.1109/TLT.2025.3587488
Ecem Kopuz;Galip Kartal
This study investigates how educational researchers integrate artificial intelligence (AI) tools into their workflows, with a focus on balancing automation and human judgment. The study, which provides a mixed method approach with a survey and interview questions, utilized an international sample of 65 educational research fields. The findings reveal that AI-supported tools help reduce the burden while carrying out research processes, so that more time can be spent on basic and innovative activities. In addition, ethical and practical guidelines have emerged on how to optimize human–AI collaboration. It has been determined which tools researchers use and how. This study attempts to explain how AI can be effectively integrated with human intelligence. Considering this, it emphasizes the need to create strong policies and standards on the use of AI, to raise awareness of users about technology use, and to ensure that ethical practices are observed. This article offers a roadmap outlining which AI tools can be used and in what ways. It also makes significant contributions to the literature in this field by emphasizing the indispensable importance of human intervention in intelligence-supported education research.
{"title":"Collaborative Human–AI Research Practices: Identifying Critical Touchpoints for Human Intervention in Educational Research","authors":"Ecem Kopuz;Galip Kartal","doi":"10.1109/TLT.2025.3587488","DOIUrl":"https://doi.org/10.1109/TLT.2025.3587488","url":null,"abstract":"This study investigates how educational researchers integrate artificial intelligence (AI) tools into their workflows, with a focus on balancing automation and human judgment. The study, which provides a mixed method approach with a survey and interview questions, utilized an international sample of 65 educational research fields. The findings reveal that AI-supported tools help reduce the burden while carrying out research processes, so that more time can be spent on basic and innovative activities. In addition, ethical and practical guidelines have emerged on how to optimize human–AI collaboration. It has been determined which tools researchers use and how. This study attempts to explain how AI can be effectively integrated with human intelligence. Considering this, it emphasizes the need to create strong policies and standards on the use of AI, to raise awareness of users about technology use, and to ensure that ethical practices are observed. This article offers a roadmap outlining which AI tools can be used and in what ways. It also makes significant contributions to the literature in this field by emphasizing the indispensable importance of human intervention in intelligence-supported education research.","PeriodicalId":49191,"journal":{"name":"IEEE Transactions on Learning Technologies","volume":"18 ","pages":"732-740"},"PeriodicalIF":2.9,"publicationDate":"2025-07-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144680846","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This study verifies the ability of large language models (LLMs) to generate a curriculum and develop syllabi for specific courses. We prompted four models to generate two sets of curricula for a bachelor’s degree in Economics and Management. We also generated syllabi for the courses included in the curriculum. We chose five Polish public economics universities offering those degree programs for comparison. Four LLMs were used in this experiment: ChatGPT-3.5, ChatGPT-4, Google Bard, and Gemini. Two of them are multimodal models. The study used an iterative approach, increasing the detail of the prompt in each iteration. The results show that the more specific prompt is given to the LLM, the less accurate the results are. Moreover, the experiment shows that none of the LLMs developed a complete curriculum at a level comparable to that generated by humans. However, LLMs can significantly help create a curriculum and develop syllabi by humans, provided that there is close human–artificial intelligence (AI) collaboration. The results obtained from the AI-assisted curriculum design differ depending on the model. By analyzing the differences between the tools and the real degree programs and syllabi, we determined that multimodal models are better suited for this task than older models.
{"title":"Generative AI in Curriculum Design: Empirical Insights Into Model Performance and Educational Constraints","authors":"Paulina Rutecka;Karina Cicha;Mariia Rizun;Artur Strzelecki","doi":"10.1109/TLT.2025.3587081","DOIUrl":"https://doi.org/10.1109/TLT.2025.3587081","url":null,"abstract":"This study verifies the ability of large language models (LLMs) to generate a curriculum and develop syllabi for specific courses. We prompted four models to generate two sets of curricula for a bachelor’s degree in Economics and Management. We also generated syllabi for the courses included in the curriculum. We chose five Polish public economics universities offering those degree programs for comparison. Four LLMs were used in this experiment: ChatGPT-3.5, ChatGPT-4, Google Bard, and Gemini. Two of them are multimodal models. The study used an iterative approach, increasing the detail of the prompt in each iteration. The results show that the more specific prompt is given to the LLM, the less accurate the results are. Moreover, the experiment shows that none of the LLMs developed a complete curriculum at a level comparable to that generated by humans. However, LLMs can significantly help create a curriculum and develop syllabi by humans, provided that there is close human–artificial intelligence (AI) collaboration. The results obtained from the AI-assisted curriculum design differ depending on the model. By analyzing the differences between the tools and the real degree programs and syllabi, we determined that multimodal models are better suited for this task than older models.","PeriodicalId":49191,"journal":{"name":"IEEE Transactions on Learning Technologies","volume":"18 ","pages":"757-768"},"PeriodicalIF":4.9,"publicationDate":"2025-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144750849","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-07-01DOI: 10.1109/TLT.2025.3575030
May Hung May Cheng;Zhi Hong Wan
{"title":"Science Education in the Age of Artificial Intelligence: Opportunities, Challenges, and Research","authors":"May Hung May Cheng;Zhi Hong Wan","doi":"10.1109/TLT.2025.3575030","DOIUrl":"https://doi.org/10.1109/TLT.2025.3575030","url":null,"abstract":"","PeriodicalId":49191,"journal":{"name":"IEEE Transactions on Learning Technologies","volume":"18 ","pages":"635-638"},"PeriodicalIF":2.9,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11062459","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144524382","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：