Select the full screen icon to view the infographic in more detail.
Active feedback using Chat GPT
BILT Education Development Project 23/24
Sam Bell University of Bristol Business School
Project Brief
Participation in a teaching innovation project led by Jenni Rose (University of Manchester) exploring how Generative AI chatbots such as Chat GPT can facilitate ‘active feedback’ for students on their academic writing assessments. This project was part of BILT’s Artificial Intelligence and Assessment theme.
What is Active Feedback?
Students are empowered to generate ‘inner’ feedback themselves by comparing a piece of their work to comparators. Questions are provided to students to generate their own feedback from the comparison process. Students are encouraged to write down their feedback and use it to inform their next piece of work (Nichol 2022). Click this link to watch Jenni Rose from the University of Manchester introduce Active Feedback.
Can Chat GPT facilitate active feedback?
Can ChatGPT Facilitate Active Feedback?
There are two ways Generative AI chatbots can assist in the active feedback process:
- Comparators can be generated using Generative AI and/or
- Generative AI can be prompted by students to act as a facilitator of comparison in the active feedback process
Active Feedback Scaffolding Activities:
- Develop or source comparators (e.g. coursework brief, marking rubric, exemplar essay, articles, reports)
- Interactive teaching session:
- Introduces active feedback
- Enables students to develop chatbot prompts to generate feedback by comparing their work to the comparator(s)
- Unsure about prompts? See this resource from Monash University: Engaging with AI tools through prompting
- Experiments using a short piece of writing and a provided comparator
- Uses a set of questions to encourage reflection and generate feedback – some examples are covered in this Times Higher Education article
- Encourages students to write down their feedback and consider how they will use it to improve their work next time (feed forward).
Comparator Examples:
| Student work | Comparator |
| Essay or project plan | Coursework brief |
| Draft paragraph | Authoritative published sourceAcademic writing advice/model |
| Draft essay or project | Marking rubricExemplar essay (need not be on the same topic)Chat GPT generated essay |
Findings and Impact:
Thematic analysis of student reflections and focus groups across two final-year undergraduate units found that many students use Generative AI tools for feedback on their work, although not necessarily using the active feedback approach. In focus groups, students highlighted problems with Generative AI feedback, finding it too positive – highlighting the importance of prompt craft and reflection in the active feedback process. Some students expressed a preference for on-demand Generative AI feedback over lecturer feedback for reasons of confidence (“I can ask it things I wouldn’t ask my teachers”) and staff availability.
Next Steps and Contact:
My next steps will be to embed active feedback in formative assessments in 2024/2025. Although it is difficult to monitor student outcomes (as the nature of assessment is changing), I plan to gather more student opinions on the effectiveness of this approach.
Contact: Sam Bell, Senior Lecturer, University of Bristol Business School
To find out more about Active Feedback, contact Jenni Rose (University of Manchester)
Leave a Reply