I Use AI. I Also Catch My Students Lying About It!

AI in education has the potential to be a genuinely useful tool. It can enhance and speed up our work, but it should never replace us.

I've worked as an online teacher in a virtual classroom since 2007, and I can't count the number of arguments I've had with students insisting their work "wasn't done by AI" when it very clearly was. Come on, how does a student go from barely stringing a sentence together to suddenly sounding more eloquent and articulate than I am? It reminds me of my days teaching in a physical classroom, when students would swear blind they weren't on their phones, and no, they don't even own one. It creates a tense, often contentious atmosphere, when all any teacher wants to do is help their students learn. To make matters worse, my head of department has told us to simply accept the work as the student's own and not challenge them on it. I've even had one or two students pulled from my teaching groups and moved elsewhere after I dared to raise the issue.

And yes, I use AI myself. When and why? When I'm short on time and need help marking three classes' worth of Year 11 review work. I won't pretend I don't feel a twinge of guilt about it, like I'm cheating somehow. Do I check the AI's marks? Do I make sure I agree with its comments? Always. I've also used it to help plan the odd tutoring session,  though never for my GCSE students, where I stick closely to past exam papers rather than AI-generated material. I've also used it to "sharpen up" the reports I write for every individual tutee. I see these as an essential part of my role — parents are trusting me with their child for an hour and paying me for it, and they deserve to know what we actually covered in the session. I know from my own children how little you get out of "what did you do today?" and "was it okay?", so a personalised, individual report matters. AI's role there is simply to sharpen my wording, especially after a long day of back-to-back teaching and tutoring. For me, AI is only ever an enhancement. It will never replace me or the expertise I've built in this job — expertise that, ironically, I owe in large part to years of marking for AQA.

Sitting down this June to mark 300 English Language scripts for an exam board, I found myself thinking: my grandchildren will not be doing this job. There's no way on earth it survives that long. Marking English Language is trickier than a subject like maths, the criteria are notoriously fine-grained. What separates a "simple" comment from "some" comment? What counts as communicating with "some success" versus "clear" success? What tips a response into "perceptive," or "convincing," or "compelling"? You fall into such a rigid rhythm marking these scripts that I'm genuinely surprised humans are still doing it. I think it's a job AI could absolutely be trained to do, and honestly, I'd welcome it. The marking schedule is gruelling enough on its own, without also being a full-time teacher racing against the same deadlines.

AI reminds me of autocorrect, in a way. Do you remember when autocorrect first appeared? Have you gotten lazier with your spelling, trusting your phone or computer to catch your mistakes? I certainly have. I worry AI will eventually reach a point where neither teachers nor students have to think for themselves at all. That prospect genuinely frightens me.