Editorial
The landscape of education and…
by Dr Rachel Goh, Senior Specialist, ELIS & Lynn Loy, Lead Teacher English, Pasir Ris Secondary School
As English Language (EL) teachers, we observed that students often fail to recognise recurring issues in their writing such as common grammatical mistakes, unclear thesis statements and weak paragraph structures. As a professional learning team (PLT), we agreed that our students require targeted guidance to improve their writing. We were interested to explore how GenAI feedback could help students write more independently. So, in our PLT, we inquired into the following research question, ‘How can the integration of GenAI feedback practices into EL writing instruction help students plan, monitor, and revise their writing?’
In our inquiry, we employed the use of Scribo Writing Check and the built-in Claro chatbot in our writing lessons. We collaboratively planned a series of writing lessons for enactment. We began with post-writing activities involving responses to Writing Check feedback on language and grammar aspects in the first lesson study cycle. Then, we worked with pre-writing activities involving feedback from the chatbot on initial ideas in the second lesson study cycle. We collected and analysed students’ writing, history of revisions and chatlogs to identify improvement and gaps. A post-lesson whole-class survey and focus group discussions (FGD) with selected case pupils were carried out to understand their experiences.
Here are some things we learnt from our study:
- Students generally found feedback in the form of one suggestion per paragraph (Figure 1) more helpful than Glow and Grow comments that focused on what they did well in and areas for improvement. Some learners struggled with the cognitive load of long feedback comments and preferred more specific suggestions.
Figure 1. Feedback in the form of one suggestion per paragraph
(2) When interacting with the chatbot at the pre-writing stage, students had better outcomes when they were guided to prompt the bot to explain its reasoning (Figure 2).
Figure 2. Student’s prompting the chatbot to explain its reasoning
(3) Referencing the history of revisions (Figure 3) was helpful for students to track their thinking as they engaged with feedback generated by Scribo Writing Check. They became more aware of their areas for improvement in the next piece of writing.
Figure 3. History of revisions
When we analysed the focus group discussions, we could categorise students into three broad categories: Passive, Active and Strategic, based on their experiences with GenAI feedback.
Passive Engagement (Least desirable)
Students who demonstrated passive engagement had viewed that GenAI feedback was for Correction and Improvement. Their focus was on following what GenAI told them to do because it knew better than they did. They would accept the feedback readily believing that the suggestions would improve their work. The concern here is that the students are uncritical in accepting feedback. This uncritical stance is evident when students position GenAI feedback as instructions to be followed rather than suggestions to be reviewed, resulting in minimal evaluation of whether feedback aligns with meaning, context or task requirements. This orientation was illustrated by Student X’s survey responses: ‘I just asked the AI and changed each mistake accordingly … I straight away copied it down without reading it … I ignored some of the suggestions as I asked the questions wrong myself.’ Figure 4 shows the history of revisions from Student X’s written speech shows how he passively engaged with GenAI feedback and how that is not ideal.
Figure 4. Student X’s history of revisions
Student X uncritically accepted the GenAI suggestion to replace the coordinating conjunction ‘and’ with the conjunctive adverb, ‘furthermore’ that acts like a signpost to tell the reader how the second thought relates to the first, in this case to add information. Doing so made the reading of the revised speech sound disjointed.
Active Engagement
Students who demonstrated active engagement had viewed that GenAI feedback was for Consideration and Negotiation. Their focus was on evaluating if the feedback fit what they were trying to write and the writing task. They would assess each suggestion to see if it made sense for their writing. Student ZY’s survey responses illustrated this orientation: ‘I used most of the feedback given by Scribo to help shorten my sentences or combine multiple into one… I ignore some as the suggestions did not meet my context and in turn did not make much sense.’ The following shows an excerpt of ZY’s original writing (Figure 5) and the chatbot’s feedback (Figure 6). The revised writing (Figure 7) shows how the student had actively engaged with GenAI feedback by adapting the ideas to his specific context.
While active engagement indicates increasing discernment, it is only when students begin to plan how, when and why to use GenAI alongside other feedback sources that deeper learning and transfer can occur.
Strategic Engagement (most desirable)
Students who demonstrated strategic engagement had viewed that GenAI Feedback was for Orchestration. Their focus was on using GenAI as part of learning to get better at writing. Student Z demonstrated such an orientation towards orchestration, as illustrated in her survey responses: ‘When I received the AI-generated feedback, I first read through all the comments carefully to understand the suggestions, then identified which ones addressed actual errors such as grammar or cohesion. Next, I decided which stylistic suggestions were unnecessary and set them aside, and finally, I revised my essay by correcting the errors, adding connectors, and improving clarity where needed.’
Students who demonstrated strategic engagement would address GenAI feedback as an option, making sense of it before planning when and how to use it along with other feedback sources. Figure 8 shows what Student JE had drawn during the FGD as a visual representation of how she unpacked the writing requirements at the pre-writing stage, engaged with GenAI while-writing to refine initial ideas and elicited feedback from her teacher after revising writing.
Figure 8. Student JE’s visual representation of how she engaged with GenAI feedback and teacher feedback
The three different ways students engaged with GenAI feedback and the implications for teacher practice is summarised in Figure 9 below.
When teachers notice that their students are simply copying GenAI suggestions without thinking (passive engagement) or only accepting feedback that seems helpful (active engagement), targeted instruction can be used to support students in moving towards more strategic engagement with feedback. For students demonstrating passive engagement (correction-seeking or improvement-focused), teach them to pause and ask, “Why might this suggestion work better?” before making changes to their writing. Encourage them to compare their original writing with GenAI suggestions and evaluate whether the feedback fits their intended message and audience. For students at the active engagement level (consideration-based or negotiation-oriented), have them reflect on the criteria they had used to decide which GenAI suggestions to accept or reject so that they become conscious of their evaluation strategies rather than make intuitive decisions without understanding why. Encouraging them to reflect on their learning as a writer shifts the focus from producing one good piece of writing to growing their writing capabilities and independence. For students who are already strategic in their feedback engagement, they benefit from opportunities to apply their sophisticated approach to different text types and contexts—from school writing to writing for authentic real-world situations —helping them adapt their systematic feedback processes to new writing demands and audiences. Collectively, these instructional moves support teachers in responding to passive and active engagement while intentionally guiding students towards strategic, transferable use of GenAI feedback.
Our realisations:
With the integration of GenAI feedback in writing instruction, we experienced a shift towards feedback provision becoming less teacher-driven and more co-constructed with students taking greater ownership in engaging the chatbot for feedback and interpreting and acting on feedback generated by the Writing Check. Our experiences find resonance with the idea of shared responsibilities in feedback processes (Nash & Winstone, 2017) where roles and responsibilities are co-constructed with students to help them understand that while teachers are responsible for designing feedback activities, they bear the responsibility to seek and clarify feedback and use it for learning.
Teachers’ reflections further highlighted the importance of pedagogical mediation in the use of GenAI feedback. While tools such as Scribo and Claro provided timely and visual feedback that helped students identify language and structural issues independently, teachers observed that the quality of learning depended heavily on how students were guided to interpret, prioritise and act on this feedback. For lower-progress learners, unrestricted feedback could feel overwhelming, requiring teachers to reduce cognitive load and model how to focus on key areas for improvement. Importantly, teachers noted that when students were explicitly guided to evaluate and negotiate GenAI feedback, they were more likely to move from passive acceptance towards more strategic engagement over time. Across classes, GenAI feedback also enabled teachers to shift their own focus away from surface-level correction towards supporting idea development, coherence and voice. These reflections reinforce that while GenAI can enhance feedback processes, it is teachers’ professional judgement, scaffolding and relational support that enable feedback to translate into meaningful learning. This underscores that GenAI feedback does not diminish the teacher’s role, but rather amplifies the need for intentional scaffolding, modelling and dialogic feedback to support students’ evaluative and metacognitive growth.
Drawing on our experiences, we represented our thinking in an infographic about the teacher’s role and the students’ role when engaging with GenAI feedback in the writing process. This infographic which was included in the EL Secondary Quick Guide to GenAI (p. 6) would inform our future practice. Our realisations from the inquiry are about the importance of positioning GenAI feedback in ways that amplify student voice and not override it as well as supporting writers’ growth by helping them maintain ownership of their writing as they work with their initial ideas to the revision of writing.
References
Nash, R. A., & Winstone, N. E. (2017). Responsibility-sharing in the giving and receiving of assessment feedback. Frontiers in psychology, 8, 1519.