Search

Assessment feedback is important to student learning. Learning analytics (LA) powered by artificial intelligence exhibits profound potential in helping instructors with the laborious provision of feedback. Inspired by the recent advancements made by Generative Pre-trained Transformer (GPT) models, we conducted a study to examine the extent to which GPT models hold the potential to advance the existing knowledge of LA-supported feedback systems towards improving the efficiency of feedback provision. Therefore, our study explored the ability of two versions of GPT models – i.e., GPT-3.5 (ChatGPT) and GPT-4 – to generate assessment feedback on students' writing assessment tasks, common in higher education, with open-ended topics for a data science-related course. We compared the feedback generated by GPT models (namely GPT-3.5 and GPT-4) with the feedback provided by human instructors in terms of readability, effectiveness (content containing effective feedback components), and reliability (correct assessment on student performance). Results showed that (1) both GPT-3.5 and GPT-4 were able to generate more readable feedback with greater consistency than human instructors, (2) GPT-4 outperformed GPT-3.5 and human instructors in providing feedback containing information about effective feedback dimensions, including feeding-up, feeding-forward, process level, and self-regulation level, and (3) GPT-4 demonstrated higher reliability of feedback compared to GPT-3.5. Based on our findings, we discussed the potential opportunities and challenges of utilising GPT models in assessment feedback generation.

In higher education, delivering effective feedback is pivotal for enhancing student learning but remains challenging due to the scale and diversity of student populations. Learner-centered feedback, a robust approach to effective feedback that tailors to individual student needs, encompasses three key dimensions—Future Impact, Sensemaking, and Agency, which collectively include eight specific components, thereby enhancing its relevance and impact in the learning process. However, providing consistent and effective learner-centered feedback at scale is challenging for educators. This study addresses this challenge by automating the analysis of feedback content to promote effective learner-centered feedback principles. We gathered a large-scale dataset of 16,531 feedback entries from 95 courses from one Master's and one Bachelor's program within a large faculty at an Australian university, with each entry labeled in accordance with the learner-centered feedback framework. Employing a range of machine learning and deep learning techniques, including Random Forest, XGBoost, BERT, and GPT-3.5, we systematically investigated the effectiveness of different approaches for constructing classifiers to accurately categorize feedback into various learner-centered components. The results demonstrated that the BERT-based classifiers outperformed other models in most feedback categories (achieving Cohen's κ up to 0.956 and F1 score up to 0.998), but showed relatively low performance in categories with less training data. This automated analysis aids in scrutinizing feedback quality, thereby supporting educators in enhancing their feedback practices to be more aligned with learner-centered principles.

There are emerging concerns about the Fairness, Accountability, Transparency, and Ethics (FATE) of educational interventions supported by the use of Artificial Intelligence (AI) algorithms. One of the emerging methods for increasing trust in AI systems is to use eXplainable AI (XAI), which promotes the use of methods that produce transparent explanations and reasons for decisions AI systems make. Considering the existing literature on XAI, this paper argues that XAI in education has commonalities with the broader use of AI but also has distinctive needs. Accordingly, we first present a framework, referred to as XAI-ED, that considers six key aspects in relation to explainability for studying, designing and developing educational AI tools. These key aspects focus on the stakeholders, benefits, approaches for presenting explanations, widely used classes of AI models, human-centred designs of the AI interfaces and potential pitfalls of providing explanations within education. We then present four comprehensive case studies that illustrate the application of XAI-ED in four different educational AI tools. The paper concludes by discussing opportunities, challenges and future research needs for the effective incorporation of XAI in education.

Fast improvements in computing power and Artificial Intelligence (AI) algorithms enable us to automate important decisions that shape our everyday lives, and drive workplace transformations. It is predicted that many people will find themselves unprepared to deal with high degrees of change and uncertainty, increasingly posed by AI in some sectors. A critical educational challenge involves figuring out how to support young generations to develop the capabilities that they will need to adapt to, and innovate in, a world with AI. This article argues that both educators and learners should be involved not only in learning but also in co-designing for learning in an AI world. Further, they together should explore the knowledge, goals and actions that could help people shape future AI scenarios, and learn to deal with high degrees of uncertainty. A key contribution of the paper is a re-conceptualization of design for learning in an AI world, which explores a problem space of educational design, and illustrates how educators and learners can work together to re-imagine education futures in an AI world. As part of this problem space, the paper discusses underpinning philosophies (the capability approach and value creation), a high-level pedagogy (with an emphasis on co-creation), pedagogical strategies (speculative pedagogies), and pedagogical tactics (AI scenarios). It then proposes a design framework (ACAD) to support educators and learners' discussions about design for learning in an AI world. This participatory design approach aims to sensitize people for what education may mean, for whom, and how learning with AI may look like, and it highlights the active engagement of educators and learners in co-designing a future they desire, to help shape learning and living in an AI world.

Feedback is an essential component of scaffolding for learning. Feedback provides insights into the assistance of learners in terms of achieving learning goals and improving self-regulated skills. In online courses, feedback becomes even more critical since instructors and students are separated geographically and physically. In this context, feedback allows the instructor to customize learning content according to the students' needs. However, giving feedback is a challenging task for instructors, especially in contexts of large cohorts. As a result, several automatic feedback systems have been proposed to reduce the workload on the part of the instructor. Although these systems have started gaining research attention, there have been limited studies that systematically analyze the progress achieved so far as reported in the literature. Thus, this article presents a systematic literature review on automatic feedback generation in learning management systems. The main findings of this review are: (1) 65.07% of the studies demonstrate that automatic feedback increases student performance in activities; (2) 46.03% of the studies demonstrated that there is no evidence that automatic feedback eases instructors’ workload; (3) 82.53% of the studies showed that there is no evidence that manual feedback is more efficient than automatic feedback; and (4) the main method used for automatic feedback provision is the comparison with a desired answer in some subject (such as logic circuits or programming).