Abstract:

There is ongoing interest in how generative AI tools, particularly large language models, can be effectively integrated into learning management systems and online education programmes. As feedback plays a critical role in learning and assessment, educators must carefully consider how AI can be used to support students in meaningful and pedagogically sound ways. While much existing research has focused on comparing the quality of AI-generated feedback with human feedback, this paper explores an alternative approach by proposing a personalised feedback framework that extends the capabilities of ChatGPT through the integration of embeddings and retrieval mechanisms.

The framework enables a more informed understanding of course materials and supports topic-targeted feedback for assessment activities, particularly quizzes. We describe the design and implementation of the framework within an online cybersecurity programme and evaluate its performance across both open-ended and multiple-choice questions. The evaluation shows efficacy rates of approximately 90% for free-text questions and 100% for multiple-choice questions, demonstrating that the proposed approach can provide timely, coherent, and context-aware feedback. These findings suggest that generative AI, when combined with structured course content and retrieval techniques, has significant potential to enhance feedback practices in online computing education.

Keywords:

Computing education, personalised feedback, generative AI, learning management systems, online assessment, cybersecurity education.