Closing the Feedback Gap with AI Supported Reflection
Background Information
Reflection and self-assessment are increasingly used in IE&IS courses to support student motivation, metacognitive development, and growth mindset. In several courses, students regularly reflect on their learning process using structured, process based rubrics. These activities help students make sense of their development, but they also rely heavily on (formative) feedback from teachers to be effective.
Providing high quality feedback on reflections is time intensive. As student numbers grow, teachers face increasing pressure to balance feedback quality with workload. This limits scalability and can delay feedback, reducing its impact on student learning and self-regulation. As a result, students may not receive timely guidance when they need it most, while teachers struggle to maintain consistency across and within cohorts.
At the same time, developments in large language models offer new opportunities to support reflective learning. While many educational pilots focus on AI for assessment or question answering, fewer address the depth of process based reflection and feedback. This pilot responds to that gap by exploring how AI can be embedded directly into reflective learning activities, without replacing teachers, and while respecting privacy and pedagogical values.
By focusing on formative feedback for reflection rather than grading, the project addresses a core challenge in contemporary higher education: how to support deep learning and self-directed development at scale.
Aim of the project
The aim of this project is to embed AI-based chatbots into course-level reflection activities in order to provide students with immediate, personalized, and rubric-aligned formative feedback. The project focuses on supporting the learning process rather than evaluating outcomes, helping students improve goal-setting, self-monitoring, and self-evaluation skills.
For students, the project aims to strengthen metacognition, motivation, and ownership of learning by ensuring that feedback is timely, actionable, and aligned with clear developmental criteria. Instead of waiting for teacher comments, students receive guidance during the reflection process itself, allowing them to adjust and deepen their thinking while learning is still ongoing.
For teachers, the project aims to reduce repetitive feedback workload while increasing insight into student learning. Aggregated analytics dashboards will highlight common gaps and trends in student reflections, enabling instructors to adapt teaching strategies, target coaching efforts, and improve course design.
At an institutional level, the project aims to develop a scalable and responsible model for using AI in reflective learning cycles. Through a human-in-the-loop approach, continuous calibration, and privacy-conscious implementation, the project seeks to generate practical guidelines and evidence to support broader adoption of AI-supported feedback across TU e and beyond.
For more information, please contact:
