AI in education: research into the Feedback Companion at Open-ICT

The role of AI in education is developing at lightning speed. Where students used to depend on lecturers and fellow students for feedback, AI tools now make it possible to receive immediate and interim feedback. This can not only speed up the learning process, but also contribute to a better understanding of learning goals and self-regulation.

Imagine: as a student you are working on an assignment and immediately receive clear feedback on your outline. No long waiting times, no endless searching through study guides – just directly applicable suggestions, so your work improves step by step. This scenario is not some distant prospect, but reality with AI-supported feedback tools.

Recent research, “Transforming Higher Education with AI: A Study on Engagement, Learning Outcomes, and Ethical Implications”, shows that AI-supported feedback has a positive effect on learning outcomes and student satisfaction. Students who use AI feedback improve their work considerably faster and experience fewer barriers to asking for feedback.

At Open-ICT at Hogeschool Utrecht we are going to examine this development critically. In the second year of the programme we are starting research into the use of the Feedback Companion: an AI-driven tool that supports students in their learning process by giving immediate feedback on their assignments and progress.

What exactly is Open-ICT researching?

Our focus is not only on the effectiveness of AI-driven feedback, but also on the broader question: does the Feedback Companion contribute to the development of feedback literacy? In other words, does the tool help students to better understand, process and use feedback in their learning process?

To investigate this, we look at:

• The quality and applicability of AI feedback: To what extent does the feedback help students improve their work?
• The impact on self-regulation and ownership: Do students use the feedback to actively steer their learning process?
• The role of AI feedback alongside lecturer feedback: How does AI feedback relate to traditional feedback from lecturers and peer feedback?
• The experience of students and lecturers: Is the feedback experienced as useful and motivating?

By investigating these aspects, we want to map objectively whether and how AI can play a role in strengthening the feedback culture within Open-ICT.

The Feedback Companion as a means

The Feedback Companion is an AI-driven feedback tool that helps students improve their (professional) products. The tool provides immediate and interim feedback based on the submitted (professional) products and the linked learning outcomes (skills and professional tasks). This means that students may be able to:

• Gain faster insight into the strengths and areas for improvement in their work;
• Process feedback at the moment it is needed, instead of days or weeks later;
• Strengthen self-regulated learning, by engaging more actively with their learning process;
• Experience lower barriers to asking for feedback, which can increase engagement and motivation.

For lecturers and students, this can mean a more impactful dialogue!

Why research into feedback literacy?

Feedback is essential for learning, but receiving feedback does not automatically mean that students know how to deal with it well. Feedback literacy is about understanding, processing and applying feedback. Students who are skilled at this can learn more effectively and improve their work in a more targeted way.

With this research we want to gain insight into the question whether AI-supported feedback helps to strengthen this skill. Can students reflect better on their work thanks to the Feedback Companion? Do they use feedback more effectively? And how does this affect their learning process and final results? What is the impact on the dialogue with a ‘peer feedbacker’ and the lecturer who is their coach?

A step forward, with an eye for quality and ethics

Although AI offers new possibilities, at Open-ICT we consider it important to approach this development critically and responsibly. That is why the Feedback Companion is being tested in a controlled and secure environment. The gpt-4o model runs via Microsoft Azure. Students and lecturers gain insight into how the tool works and which data is used, so that transparency and privacy remain guaranteed.

Moreover, we do not see AI as a replacement for lecturers, but as a supporting technology. Ultimately, human interaction and guidance remain essential in education. This research helps us understand how AI feedback can be a valuable addition without undermining the role of lecturers.

What does this mean for the future?

This research at Open-ICT is a first step in critically exploring the role of AI in feedback processes. The insights will help us determine whether and how AI feedback can make a positive contribution to the learning experience of students.

The results can yield valuable insights not only within Open-ICT, but also in a broader educational context. The better we understand how AI can support feedback, the more effectively we can use this technology to improve education.

Open-source availability of the Feedback Companion

An important principle in the development of the Feedback Companion is that educational innovation should be accessible and transparent. That is why we are going to make the Feedback Companion available as open source. This means that anyone – from educational institutions to individual teachers and developers – can download, adapt and implement the tool to suit their own needs.

The code will be freely accessible via GitHub, so that schools and universities can integrate the Feedback Companion into their own teaching practice. Through its open-source nature, we want to encourage a community of educational developers, teachers and researchers to work together on improvements, extensions and integrations with existing learning platforms.

With this step we ensure that the technology remains transparent, verifiable and adaptable. It also gives educational institutions the option to host the Feedback Companion within their own infrastructure, so that privacy and data security are entirely in their own hands.

The GitHub repository will be available soon, together with documentation and installation guides. Keep an eye on Education Oracle and our channels for the official release!