AI and the struggle in learning

A newly arrived student, aged 18, enrolled in the school in the second half of the year. For ease of discussion, I am going to call this student Mali. The family was educated, in their own language, and had some English language ability. My sampling of work shortly after enrolment placed her at a bandscale 3 in speaking, listening and reading and just below in writing. She was highly driven and keen on getting into senior schooling the next year. Her language class was working on assessments at the time, and a couple of days before the draft was due, she brought her draft to me and asked for feedback.

I had one glance through the paper and immediately recognised it was AI generated. I looked at Mali and directly asked her – is this written by AI? Her reply didn’t surprise me as much as her surprised look as she saw my disappointment clearly written on my face. Her answer – “can’t I do that?” – floored me. I asked her which part of the paper was written by her. Her answer – she’d input the question and whatever she understood from her teacher’s explanation. Now this assessment called for an analysis of a satirical image. I noticed nothing had been included in the paper about the image, and she responded that she didn’t know much about the image.

There is a lot to unpack in that one scenario. At this point, I’d like to focus on the use of AI. Do I believe that our students should be using AI? Absolutely! The reality is that it is in our lives and hiding our head under the sand isn’t going to make it any easier for our students. The reality is that when students are in high school, keen to reach their personal goals, they will use any resource available to them. It is critical therefore that we acknowledge this and prepare for its use in the classroom and for students’ learning.

A recent ASCD blog entry by Joseph South explores the difference between productive and unproductive struggle. South talks about the repetitive work that AI can help take on so that the actual intellectual work can be addressed. The challenge in high school is that there is value in having students learn all aspects of putting the task together, especially when they are EAL/D learners new to the whole educational set up. So, when developing the learning goals for such students, how do we identify the aspects of the task that “adds no value to the intellectual outcomes we are pursuing”¹.When students face the pressure to meet their curriculum needs, there is the temptation to use AI. This is all the more evident amongst EAL/D learners who face the barrier of language to meaning making.

When we consider the development of these learners skills, we need to have a projection of goals in mind. There are the long term goal, and short term goals. The problem for Mali is that she saw the long term goal – to submit the assessment as a finished product to be graded, as the one to focus on – largely because the whole class was focussed on that . The language conventions, stylistic choices and content demonstrating the mastery of cognitive skills, all had to be packaged neatly for submission. The reality is that these require the building of skills that will take time to develop. She will learn the language over time. So I would say that the actual writing of the task, with emphasis on language choices and literacy competence may be where she could use AI support. To ensure there is no unhealthy reliance on AI moving forward, it is imperative that students are aware of its place within their steps to success. Students must be aware that the productive struggle of creating their own text must be met eventually as they build their language competency. At the moment however, for Mali, focussing on developing this analytical essay would have been unproductive struggle.

When Mali showed me her draft, my immediate response, to support her, was to ask her for her notes. I expected to see a scaffold of some sort that would have captured how the image was broken down to its parts to enable the analysis. This scaffold of learning would have captured the cognitive skills required to analyse the image – something accessible for a student who had prior education under her belt. However, she had no notes. As she had recently arrived, that part of the lesson – the actual breaking down of the texts to make the analysis was missing. It was therefore very clear why there were no references to the picture in her AI assisted draft. She hadn’t actually analysed the image.

The EAL/D learner in a mainstream classroom can access the cognitions and concepts explored in the classroom. A multitude of tools such as translations, peers interpretations, images, can be used to get to the meaning of the texts in front of them. Mali needed to be given this opportunity to make this meaning through explicit teaching – and building this up should have been the priority in her lessons at the time. She could have been graded on her ability to demonstrate the required cognitions. This is where the productive struggle should have centred.

For Mali, at this point, trying to write a 600 word analytical essay was unproductive, and AI could be used to generate this product. It is imperative however, that an EAL/D learner should know why she is using AI at this point, and how she can, over time, rely on it less and use it more efficiently. It is imperative that the teacher knows that based on Mali’s language acquisition level and someone just two weeks into the system, she will not be able to be assessed on all aspects of the assessment.

1. Joseph South “AI in Learning: Productive Support or Cognitive Crutch?” (accessed on 5th Feb 2026) https://www.ascd.org/blogs/ai-in-learning-productive-support-or-cognitive-crutch ↩︎