This full research paper presents findings from semi-structured interviews conducted with eleven K-12 teachers, mostly in STEM subjects, to investigate how a professional development (PD) program impacted their AI competencies, in terms of attitudes, perceptions and strategies of integrating Generative AI (GAI) into their classrooms. The results show that teachers improved their knowledge and awareness of different aspects of GAI, revealing five major themes: overall positive evaluations of the PD program, shifting teacher attitudes towards GAI, perceived pedagogical benefits of GAI including personalization and multimodality, implementation challenges, and commitment to fostering student AI competencies. The study makes three key contributions. First, it provides recommendations for future PD programs supporting teachers' AI competencies. Second, it captures teachers' perspectives within the Swedish context and enriches diversity of the empirical data in this field. Finally, the findings offer valuable insights for policymakers involved in strategic planning and resource allocations to support teachers' PD.

Student wellbeing is recognised as a critical factor in supporting learning outcomes. However, when teachers are assigned expanded roles as mentors to monitor and support wellbeing, they often lack the necessary support to fulfil these responsibilities effectively. This study explores the tensions and behaviours that arise when teachers implement a novel wellbeing tool designed to provide insights into students’ wellbeing. Adopting a mixed-methods approach, data were gathered through a survey and interviews with teachers (n = 39) and analysed through the lenses of Activity Theory (AT) and the Theory of Planned Behaviour (TPB). The findings reveal that integrating the wellbeing tool introduced new practices, resulting in primary and secondary tensions concerning roles, implementation, and use. The integration of TPB uncovered psychological factors influencing teachers’ intentions and behaviours, complementing the systemic insights from AT. The study contributes to the broader conversation on enhancing educational practices to support student wellbeing, fundamental to achieving positive learning outcomes.

Learning analytics (LA) is an important area of research in technology-enhanced learning that has emerged during the last decade. In earlier years, several systematic reviews have been conducted that focused on the theories behind LA or on empirical studies that utilised LA-based methods to improve learning and teaching processes in higher education. However, to date, there has been no systematic review of papers that have adopted a software perspective to report on the many forms of learning analytics software (LAS) that have been developed, despite these being used more frequently than before in higher education to support learning and teaching processes. To fill this gap, this paper presents a systematic review of LAS with the aim of critically scrutinising the ways in which the use of interactive software in real-world settings may both support students in improving their academic performance and assist teachers in various pedagogical practices. A thematic analysis of 75 articles was conducted, resulting in the identification of three categories of LAS: at-risk student identification software; self-regulation software; and collaborative learning software. For each of these categories, we analysed (i) the embedded functionality; (ii) the stakeholder (teacher and student) for which the functionality is intended; (iii) the analytical and visualisation approaches implemented; and (iv) the limitations of the software that require future attention. Based on the findings of our review, we propose future directions for the development of LAS.

Data-driven methods are being implemented in many schools around the world to improve education. In this study, two schools were studied to investigate how they implemented datadriven methods for the monitoring and improvement of the well-being of their students. These schools were part of a Swedish national program where 15 schools participated to use data on both classroom, school, and system levels for school improvement. We identified five factors that influenced the implementations, namely data collection and analysis, frequency, anonymity, involving students, and organizational changes. We conclude that continuous and frequent data collection provided insights on students´ well-being that cannot be achieved without systematic data collection. Since this kind of data collection can be time-consuming, dedicated digital tools can be used to automate data collection and analysis. These tools can also provide a better basis for decision-making since it is easier to connect and visualize data. We also conclude that the European Union’s (EU) General Data Protection Regulation (GDPR) is important when using student data, and there is a need for national guidelines on how to use data securely and efficiently in schools.

Även om forskning har bedrivits om självreglerat lärande i relation till lärandeanalys finns det fortfarande en kunskapslucka när det gäller de hinder som elever i gymnasieutbildningen möter i att reglera sitt eget lärande och hur lärandeanalys kan stödja deras självreglering. Denna artikel undersöker två frågor: 1) Vilka utmaningar upplever gymnasieelever i processen att reglera sitt eget lärande?, och 2) Vilken information och data behöver gymnasieelever för att bättre kunna reglera sitt eget lärande? Vi genomförde en studie på en medelstor gymnasieskola i Mellansverige för att bättre förstå hur dessa frågor manifesterar sig bland eleverna. Vi analyserade data som samlats in av skolan två gånger årligen mellan 2015 och 2022 och administrerade ett frågeformulär till 224 elever för att besvara forskningsfrågorna. Genom beskrivande statistik och en tematisk analys identifierar vi vanliga problem som elever stöter på samt den information som är nödvändig för att stötta självreglerat lärande. Vi diskuterar implikationerna av våra fynd för utformningen av system som förser elever med relevant data för att förbättra deras lärandeupplevelser.

While research has been conducted on self-regulated learning in relation to learning analytics, there remains a knowledge gap regarding the obstacles secondary education students face in regulating their learning and how learning analytics can support their self-regulation. This paper investigates two questions: 1) What challenges do secondary education students experience in the process of regulating their own learning?, and 2) What information and data do secondary education students need to better regulate their own learning? We conducted a study at a mid-sized upper secondary school in middle Sweden, to better understand how these issues manifest among students. We analyzed data collected by the school twice annually between 2015 and 2022, and administered a questionnaire to 224 students to answer the research questions. Through descriptive statistics and a thematic analysis, we identify prevalent problems that students encounter, as well as the necessary information that is essential for scaffolding self-regulated learning. We discuss the implications of our findings for the design of systems that provide students with relevant data to enhance their learning experiences.

Self-regulated learning is an essential skill that can help students plan, monitor, and reflect on their learning in order to achieve their learning goals. However, in situations where there is a lack of effective feedback and recommendations, it becomes challenging for students to self-regulate their learning. In this paper, we propose an explainable AI-based approach to provide automatic and intelligent feedback and recommendations that can support the self-regulation of students' learning in a data-driven manner, with the aim of improving their performance on their courses. Prior studies have predicted students' performance and have used these predicted outcomes as feedback, without explaining the reasons behind the predictions. Our proposed approach is based on an algorithm that explains the root causes behind a decline in student performance, and generates data-driven recommendations for taking appropriate actions. The proposed approach was implemented in the form of a dashboard to support self-regulation by students on a university course, and was evaluated to determine its effects on the students' academic performance. The results revealed that the dashboard significantly enhanced students' learning achievements and improved their self-regulated learning skills. Furthermore, it was found that the recommendations generated by the proposed approach positively affected students' performance and assisted them in self-regulation

Using AI for automatically linking exercises to curriculum goals can support many educational use cases and facilitate teaching-to-the-curriculum by ensuring that exercises adequately reflect and encompass the curriculum goals, ultimately enabling curriculum-based assessment. Here, we introduce this novel task and create a manually labeled dataset where two types of diagnostic tests are linked to curriculum goals for Biology G7-9 in Sweden. We cast the problem both as an information retrieval task and a multi-class text classification task and explore unsupervised approaches to both, as labeled data for such tasks is typically scarce. For the information retrieval task, we employ SOTA embedding model ADA-002 for semantic textual similarity (STS), while we prompt a large language model in the form of ChatGPT to classify diagnostic tests into curriculum goals. For both task formulations, we investigate different ways of using textbook content as a pivot and provide additional context for linking diagnostic tests to curriculum goals. We show that a combination of the two approaches in a retrieval-augmented generation model, whereby STS is used for retrieving textbook content as context to ChatGPT that then performs zero-shot classification, leads to the best classification accuracy (73.5%), outperforming both STS-based classification (67.5%) and LLM-based classification without context (71.5%). Finally, we showcase how the proposed method could be used in pedagogical practices.

Student well-being is important for inclusive societies and academic achievement. As studies have shown, well-being is associated with school success. Today, it's common for schools to use different technologies to collect and analyse digital data with the purpose of improving educational outcomes. Generally, this collected data focus on student engagement, attendance, and results, with novel advancements being aimed at supporting student well-being. In this pilot study, we analyse teachers' experiences to identify benefits and challenges during a spearhead integration of a data-driven tool for examining student well-being in upper secondary school. Using thematic analysis of teacher interview transcripts, we identified four themes: insight diversity, caring pedagogies, teacher leadership tools, and faculty transformation. The themes are discussed using the theoretical perspectives of orchestration, practice, process, and actors. Key results show that some high-value benefits teachers report on are gaining insights, saving time, and informing decision-making. The challenges include a lack of systematisation, guidance, and resources, and tensions related to defining the role and responsibilities of a teacher or mentor. We conclude that schools that work to support student well-being can benefit from the diversity of insights and practices related to the presented tool. However, an informed and systematic approach would be needed to leverage the benefit of spearhead integration. The contribution of the study is to provide insights on how a well-being tool can be used in an educational context to bring understanding of student well-being to teachers. Our results may inform decisions and guide integration and implementation practices in schools.

Providing personalized support to students during courses is essential to facilitate them in their desired learning goals and reduce the dropout rate. Although teachers can play an effective role in providing personalized support, achieving individual-level assistance for massive courses becomes challenging. To overcome this challenge, this paper proposes a transformer-based approach that first models students’ knowledge of various course concepts based on their performance in various assessed tasks. Afterwards, the students’ concept-wise knowledge level derived from the models is combined with the available course material, leading to the generation of personalized concept-wise exercises by employing fine-tuned Text-to-Text Transfer Transformer (T5) architecture. These generated exercises help students to improve their knowledge about different course concepts. The proposed approach has been evaluated with various university courses to determine its quality, utility and effects on students’ academic performance. The evaluation results revealed that teachers and students were satisfied with the quality of the generated exercises, and these were found to be helpful for students to improve their concept-wise understanding. Furthermore, the generated exercises positively impacted students’ academic performance. 

Automated Essay Scoring (AES) automatically allocates scores to essays at scale and may help teachers reduce the heavy burden during grading activities. Recently, researchers have deployed neural-based AES approaches to improve upon the state-of-the-art AES performance. These neural-based AES methods mainly take student essays as the sole input and focus on learning the relationship between student essays and essay scores through deep neural networks. However, their only product, the predicted holistic score, is far from providing adequate pedagogical information, such as automated writing evaluation (AWE). In this work, we propose Topic-aware BERT, a new method of learning relations among scores, student essays, as well as topical information in essay instructions. Beyond improving the AES benchmark performance, Topic-aware BERT can automatically retrieve key topical sentences in student essays by probing self-attention maps in intermediate layers. We evaluate the performance of Topic-aware BERT of different variants to (i) perform AES and (ii) retrieve key topical sentences using the open dataset Automated Student Assessment Prize and a manually annotated dataset. Our experiments show that Topic-aware BERT achieves a strong AES performance compared with the previous best neural-based AES methods and demonstrates effectiveness in identifying key topical sentences in argumentative essays.