AI, the Educator and the Hypocrite

I spent two days at CSE-Connect’s Advances in Teaching and Learning for Cyber Security Education this July. Good sessions. Better conversations. Best of all in the corridor (or pub). I met people I usually only see on a screen, in the same building where I usually teach.

One thing nagged at me. Every session on AI focused on staff. Creating with it. Writing with it. Saving time with it. From what I attended, the sessions avoided the student perspective.

That gap will not stay a gap for long. Two problems are heading for each other, and they are about to collide in classrooms.

First, a quality problem. Staff are using AI to build teaching material, and doing it carelessly enough that students notice.

Second, a fairness problem. The same staff are telling students they cannot use the tool themselves.

Neither problem is new on its own. What is new is both hitting the same audience, at the same time. That audience is paying close attention.

Pillar one: the low effort problem

Every academic makes the same promise at some point. I will update this next year. It gets scrawled in a margin, said out loud in a corridor, and rarely kept. Plenty of the slides still in circulation were written under that promise.

AI does not fix this habit. It disguises it. A slide left untouched for three years used to look tired, and looking tired is what got it rewritten eventually. Feed it through a model instead and it reads fresh in thirty seconds, with nothing underneath actually changed. The promise to fix it next year quietly stops needing to be kept at all.

None of this is new. Phoned-in materials have always existed. The recycled photocopy. The decade-old PowerPoint. The worksheet nobody proofread. AI has not invented low effort. It has given it a far more recognisable signature.

You know that signature already, even outside teaching. The LinkedIn post too upbeat to be human. The email from a colleague that does not quite sound like them. The photo a little too symmetrical to be real.

Ask a room of students what they think of AI and you get very different answers. Some use it for everything, without a second thought. Some refuse it outright on principle. Some refuse it on ecological grounds. A few just do not trust it.

Whichever camp they are in, they know the signature better than we do. They are immersed in an AI-driven world. The heavy users know exactly what a lazy prompt produces. Plenty of them have written one.

Aside: Students sit on a spectrum, same as staff. Some hold their own work to a high standard. Others hand in whatever the first prompt returns and think nothing of it.

This is why a bad worksheet lands harder now than a bad photocopy used to. A photocopy said, that looks a bit lazy. This says, I know exactly how you made this, and I know it took you ten minutes.

Aside: Fees make this worse. Students are paying more for their education than any generation before them. If the price has gone up, the material should not be costing less effort to make.

Aside: There is a bigger hypocrisy sitting above mine. Universities are pushing staff towards AI for one reason above all others, more output, faster, for the same pay. Ecology rarely gets a mention in that pitch…

Pillar two: the refusal problem

Here is the double standard, stated plainly. Staff can use AI to build the material a student learns from. That student cannot use it to do the work the staff member sets.

The obvious defence is that the two acts are not equivalent. Nobody is grading the staff member’s worksheet. The assignment exists to measure the student’s own thinking, not the teacher’s. Using AI to draft a handout does not corrupt that measurement the way a student using AI to write their essay would. I think that defence is basically correct.

But being correct does not make it free. Hold the line, and whether you mean to or not, you teach a second lesson alongside the first. Not that the rule protects learning, but that the rule protects whoever gets to write it. Students do not need the argument explained to feel the gap. They feel it as this: the rule applies to whoever cannot enforce it, not to whoever benefits from bending it.

This will not be true of every assessment. Some are built from the ground up to let a student use whatever tool actually helps them, and judge the outcome on its own terms. I have not seen one yet that does this in a way I would call equitable. I would like to be proved wrong. Watch this space.

Where the two meet

These are not two separate complaints. They are cause and effect.

Careless AI use is what leaves a member of staff unable to defend their own choices. The problem was never using the tool. Using it is not the sin. The problem is using it badly, and having no good answer when a student asks why the same standard does not apply to them. A teacher who uses AI well, and can explain exactly how, holds the line on student use without much trouble. A teacher who cannot explain their own use has just handed students the evidence for the opposite argument.

It is one audience, not two. The students being told no are the same students equipped to notice when the person telling them no could not defend their own use of it if asked.

Where I’ve landed

I did not start where I am now.

I began using AI for proofreading, much as I once used Grammarly. Then for brainstorming project ideas, more of them than I could generate alone. Then for validation, testing an idea before I put it in front of a real person. That last step mattered more than it looked like at the time. I did not stop needing my industry contacts and former graduates. I stopped spending their limited time on questions that did not need a human yet.

I used to think of myself as a fence rather than a scholar. My own research was never going to fill a module. What made me useful was knowing people, knowing the field, and connecting the two. AI dims that side of the job a little. Fewer quick calls to a contact for a sanity check. But it sharpens the other side considerably: spotting connections, recommending resources, surfacing options I would not have found alone. That is a trade worth making. I call it progression, not retreat.

Now I use it across the whole cycle, from first idea through to final synthesis. I have never had the time, the resources, or the expertise to do everything I want for my students. AI closes that gap faster than anything else has.

One line I still hold hard. I will not let anything carry my name without carrying my attention as well. Verification is the easy part, a quick read to check nothing is obviously wrong. What takes longer is the part after that, sitting with a draft until it sounds like someone made it on purpose, not like someone checked the output and moved on. That is where the fonts get fixed, the layout gets reworked, the flattened voice gets rewritten until a person is audibly in the room again. I spend nearly as much time putting that care into a project as I spent building it in the first place. The day I stop taking ownership of the material is the day I lose the right to ask a student to take ownership of theirs.

Students are fickle about this, and I mean that with some fondness. Put an AI-made resource in front of a lab they were never that invested in anyway, and plenty will turn their nose up before they have even opened it. Ask the same student about something they actually care about, and they will happily let AI draft half of it without a flicker of concern. The objection was never really about the tool. It was about how much they wanted to be there in the first place.

That is a different thing from the student who refuses AI on ecological or ethical grounds, and I do not want to muddy the two. These are considered positions, held consistently across whatever material they touch. The behaviour I just described, by contrast, is closer to the weather. It changes with how bored they are that week.

I know the fondness sits oddly next to what I said earlier about universities pushing this for the wrong reasons. A cynical push can still land somewhere good, and I am just less willing to give the push credit for it.

What keeps me optimistic is not an institution’s motive. It is what the tool now lets a well-intentioned educator build: material thicker, more current, and more tailored to the cohort than anything I ever built by hand.

What changes

I will put my hand up. Until now, my labs and my assessment rules did not match what I was doing myself.

AI lets me push my own teaching further. It should let me push students towards something comparable.

This is not something to get right once and leave alone. I plan to watch it closely: comments in module surveys, informal feedback in class, what external examiners say when they see the material, and wherever it shows up in the NSS. If the double standard has not shrunk a year from now, that is worth admitting in public, not just to myself.

Aside: More on the specific changes to my assessments in a follow up post later this year, once they have actually been through a trimester rather than just a plan.

Close

CSE-Connect’s event did what a good conference should. It made me ask a question I had not asked before, and it changed what I do on Monday morning.

We teach a discipline built on trust. Spot the unauthorised change. Notice the system that has quietly stopped behaving as advertised. It would be strange if the people teaching that could not hold their own materials to the same standard.

The collision I described at the start is still coming. It is not fixed yet, and it does not have to arrive the way I described it. That part is still up to us, module by module, one assessment at a time.