Detailed written feedback is often ignored, AI feedback included. ‘Feedback is of little or no value unless learners learn from it.’ Build a composite essay from the group’s recurring errors, have teams score it against criteria, add an AI audio overview, then revise their own essays.
Procedure
- Build a composite essay from the group’s recurring problems, altered enough that no one’s text is recognisable.
- Generate an audio “podcast” overview (Google NotebookLM) with the success criteria (Content, Structure, Accuracy) in the prompt.
- Teams score the essay 1–3 per criterion with descriptors, cite evidence, and write “moving forward” points.
- Play the podcast; teams add new points.
- Discuss; for homework, revise your own essay using the feedforward points.
Theory
Kerr (2020) Giving Feedback to Language Learners (Cambridge Papers in ELT): “Learning from feedback cannot be forced: the teacher’s task is to try to create the right conditions.”
Limits
Setup takes ~1 hour; fast podcast hosts suit intermediate+; privacy when uploading essays isn’t addressed.
Apply in my context (Claude)
- The composite essay alone (no AI needed) is a strong IELTS Task 2 routine. It makes feedback a task, matching Butler’s “comments over grades” and Kerr’s conditions-for-uptake point.
Connections
The case against grades · Feedback on content - make the writer feel read · Written correction is a workload and uptake problem