Enhancing the Analysis of Student-Provided Feedback through Large Language Models

Effective student feedback mechanisms are essential for continuous enhancement in engineering education. Traditional staff-student committee meetings rely on representatives and occur infrequently, potentially limiting the breadth and immediacy of student input. To address these challenges, the Mechanical Engineering programme at Northumbria University has implemented a novel approach that integrates artificial intelligence (AI) into student feedback analysis.

Our methodology involves scheduled personal tutor group meetings held monthly, during which students complete a structured reflective template. This template prompts students to evaluate their study habits, reflect on their responses to feedback, and highlight both positive experiences and areas for improvement within the programme. Unlike conventional methods that depend on a subset of students, this approach ensures comprehensive participation from the entire cohort.

The collected qualitative data is processed using Large Language Models (LLMs), to extract key themes and actionable insights efficiently. By automating the initial analysis phase, we significantly reduce the manual effort required for qualitative data processing while maintaining a high level of accuracy and consistency. This enables the programme team to identify and respond to student concerns more rapidly and effectively than through traditional committee structures.

Early findings indicate that this AI-supported method enhances both the quality and timeliness of student feedback analysis. By capturing diverse perspectives and synthesising common themes, we can implement targeted improvements with greater responsiveness. Furthermore, the integration of LLMs facilitates a data-driven approach to curriculum and support enhancements, ensuring that student voices are central to decision-making processes.

This paper will present a detailed evaluation of the impact of this initiative, including qualitative and quantitative assessments of student engagement, programme improvements, and staff efficiency. We will also discuss the potential for scaling this AI-assisted feedback system across other disciplines and institutions. By leveraging AI in educational research and administration, we demonstrate how technology can improve the student experience and streamline institutional processes in engineering education.