I’ve tried AI for things like student feedback and been very surprised. We had a 10-mark Geography assignment on Google Docs. I thought the AI would be hopeless, but I put in a list of criteria that I wanted it to mark and the comments were spot on. In fact, there were a few things that it picked up that I thought I should have focused on more. 

So, I showed it to the students who were very impressed as well. And I said to them that morally I find that a bit weird. I’m not going to use AI grading with them. I want students to know that I’m working for them in terms of building relationships. I don’t want them to think that I just go home, press a button and the work gets marked.

I feel like a really important part of being a teacher is understanding your students and how they learn. And I feel if I just pressed ‘go’ and got the comments, that over the year I wouldn’t be developing my knowledge of each individual students’ work – how they write, and where they need improvement. It would save me time, but what else am I going to do with that time that I’ve saved? The whole point is we’re trying to teach. So yeah, that manual process of doing it the slow way is helping me to gain that knowledge myself. 

But what I did say to the students is if they can get a copy of this AI grader plug-in they could install it on their machines and check for themselves before sending work to me. And that’s where I think it would save me time, because students’ work would come to me at a much better starting point, and then I could use my expertise to help them.

(Ben, Geography teacher, Brookdale High School, 30_10_24)

**

One focal point for our DP24 research project is teachers’ sense-making around the new AI tools and technologies coming into their schools. Key here is how teachers’ encounters with AI are being shaped by their own professional self-understandings and/or subjective educational theories. We are also interested in how AI might perhaps be prompting teachers to reflect on (and even modify) these understandings.

This opening reflection from a Brookdale High geography teacher with over 15-years’ experience offers a series of fascinating insights into the complexities of how teachers are making sense of AI tools within the context of their professional work. On the face of it, then, the simple ‘AI marking’ plug-in that Ben is talking about seems perfectly capable of grading work and providing the sort of feedback that a teacher might produce. As Ben concedes, on occasion the tool might even be argued to exceed what he would have done himself.

Yet, as far as Ben is concerned, this is not sufficient justification for offloading his grading responsibilities to AI. Instead, he sees marking and feedback as a key point of developing students’ trust in his professional commitment to getting to know them and supporting their development. This relates to the idea that feedback needs to be valued by students in order to be engaged with and used. 

Ben’s interpretations chime with educationalists might describe as a ‘socio-cultural’ approach to assessment and feedback – i.e. seeing feedback an active process of knowledge development that is the shared responsibility of a teacher and their students (Pryor & Crossouard 2008, Gipps 1999). This stands in contrast to standard interpretations of assessment as the one-off production of information about the strengths and weaknesses of a piece of work and how a student might improve it.  

While this latter ‘one-and-done’ approach might well support the outsourcing of marking work to AI, Ben’s professional commitment to ‘manual’ assessment clearly does not. For this teacher, at least, grading students’ work and providing feedback is not drudge work that he is eager to be relieved of having to do. Rather, he sees grading work and providing feedback is an important part of being an educator working with a group of students over a prolonged period to support their development – what educationalists have long recognised as “a prime requirement for progress in learning” (Tunstall & Gipps 1996, p.389)

Yet, this is not to say that Ben has given up on the benefits of AI grading altogether. Most interesting, perhaps, are his concluding thoughts on how this technology might be most usefully incorporated into his workflow – overseen by students rather than teachers. Here, Ben is not arguing to preserve his status as the sole judge of a student’s work. This is not a teacher who feels professionally threated by technology that might sometimes provide better feedback than his own. Instead, Ben sees the value of AI grading as a distributed resource – as an additional source of feedback for students to be using rather than an opportunity to cut back on his own grading work. All told, this might be said to be a far more progressive perspective on AI in education than Ben simply picking the tool up to use for himself.

REFERENCES

Gipps, C. (1999). Socio-cultural aspects of assessment. Review of Research in Education, 24(1), 355-392.

Pryor, J., & Crossouard, B. (2008). A socio‐cultural theorisation of formative assessment. Oxford Review of Education, 34(1), 1-20.

Tunstall, P., & Gipps, C. (1996). Teacher feedback to young children in formative assessment: A typology. British Educational Research Journal, 22(4), 389-404.