Meet the Other Phone. Only the apps you allow.

Meet the Other Phone.
Only the apps you allow.

Buy now

Please or to access all these features

AIBU?

Share your dilemmas and get honest opinions from other Mumsnetters.

Can a uni reliably say someone has used ai to create their work?

351 replies

Unissss · 29/04/2026 22:59

i personally don’t see how tbh

OP posts:
titchy · 01/05/2026 18:52

MyBraveFace · 01/05/2026 18:45

Oh, plenty of them are desperate enough if you're not worried about attending one that's prestigious or in a trendy location.

LOL. I work in one of those - we wouldn’t take you. If you had any idea about how highly regulated the sector is you’d know that really wouldn’t happen. Far too big a risk.

HollaHolla · 01/05/2026 18:54

titchy · 01/05/2026 18:52

LOL. I work in one of those - we wouldn’t take you. If you had any idea about how highly regulated the sector is you’d know that really wouldn’t happen. Far too big a risk.

Exactly! There's so much regulation, that sounds like this PP would create nine grand's worth of work, so it would be no loss...
I love the idea that any individual student could just disagree with Regs which have been through so many levels of scrutiny.

MyBraveFace · 01/05/2026 19:06

HollaHolla · 01/05/2026 18:54

Exactly! There's so much regulation, that sounds like this PP would create nine grand's worth of work, so it would be no loss...
I love the idea that any individual student could just disagree with Regs which have been through so many levels of scrutiny.

If your regulatory process is that complex, it sounds as though I'd be better off applying for a role in your regulatory policy department to help streamline it, than trying to get a place on a course.

Some complexity/multiple levels of scrutiny is inevitable, even desirable, but if it would cost you £9k to deal with a straightforward query about one clause in your regulations, you're badly overcomplicating things, which in itself creates a risk.

titchy · 01/05/2026 19:31

MyBraveFace · 01/05/2026 19:06

If your regulatory process is that complex, it sounds as though I'd be better off applying for a role in your regulatory policy department to help streamline it, than trying to get a place on a course.

Some complexity/multiple levels of scrutiny is inevitable, even desirable, but if it would cost you £9k to deal with a straightforward query about one clause in your regulations, you're badly overcomplicating things, which in itself creates a risk.

You’d need a job at our main regulator then, and another job at our other regulators.

Or go straight to the horses mouth and look for a job in the DfE and DSIT and the Home Offie and the Treasury if you want to have a say in how HE is regulated.

Feel free - most of us would be delighted to not have so many fucking hoops to jump through.

titchy · 01/05/2026 19:39

Actually ignore my posts. I’m trying to justify why it’s unreasonable for someone to object to their university carrying out routine checking to make sure the work submitted is their own. When someone doesn’t have the integrity to see that, then no amount of explaining is going to put it there.

MyBraveFace · 01/05/2026 19:42

titchy · 01/05/2026 19:39

Actually ignore my posts. I’m trying to justify why it’s unreasonable for someone to object to their university carrying out routine checking to make sure the work submitted is their own. When someone doesn’t have the integrity to see that, then no amount of explaining is going to put it there.

It isn't an objection to routine checking, it's an objection to routine checking performed using a demonstrably unfit for purpose checking tool. Surely you can see the difference?

murasaki · 01/05/2026 19:48

And its already been clarified that that is merely one stage in the checking process. Surely you can see that?

MyBraveFace · 01/05/2026 19:52

murasaki · 01/05/2026 19:48

And its already been clarified that that is merely one stage in the checking process. Surely you can see that?

If a checking tool is not reliable within a reasonable tolerance level, it has no place forming any part of the checking process.

YourShyLion · 01/05/2026 20:10

100% they can. There are programs that can tell in seconds.

MyBraveFace · 01/05/2026 20:13

This kind of muddled thinking - we know the technology makes mistakes but someone higher up has invested a ton of money in it, and no-one actually knows how to fix it, so we're going to keep using it and gaslight anyone who criticises it - is exactly how Horizon happened.

You'd have thought lessons would have been learned, but clearly not.

YourShyLion · 01/05/2026 20:14

Spell check and ai are two different things. I've never come across a spell checker that asks about rewriting into a certain style. IF what you're saying is true, which I doubt, use a better spell checker in future.

titchy · 01/05/2026 20:57

MyBraveFace · 01/05/2026 20:13

This kind of muddled thinking - we know the technology makes mistakes but someone higher up has invested a ton of money in it, and no-one actually knows how to fix it, so we're going to keep using it and gaslight anyone who criticises it - is exactly how Horizon happened.

You'd have thought lessons would have been learned, but clearly not.

Except people were convicted purely on the evidence in the Horizon system. No one is found guilty of academic misconduct purely on the basis of a Turnitin score.

Snorerephron · 01/05/2026 21:05

MyBraveFace · 01/05/2026 20:13

This kind of muddled thinking - we know the technology makes mistakes but someone higher up has invested a ton of money in it, and no-one actually knows how to fix it, so we're going to keep using it and gaslight anyone who criticises it - is exactly how Horizon happened.

You'd have thought lessons would have been learned, but clearly not.

Agree.
And I know two families whose lives were absolutely destroyed by that scandal. Unthinking trust in AI scanning software is deeply flawed

MyBraveFace · 01/05/2026 21:22

titchy · 01/05/2026 20:57

Except people were convicted purely on the evidence in the Horizon system. No one is found guilty of academic misconduct purely on the basis of a Turnitin score.

The principle is the same.

If it doesn't work and can't be fixed, it should not be used.

titchy · 01/05/2026 21:40

MyBraveFace · 01/05/2026 21:22

The principle is the same.

If it doesn't work and can't be fixed, it should not be used.

So if a lecturer ‘feels’ a student’s essay is AI written - what then? A feeling isn’t objective, so shouldn’t be used? A brilliant essay produced when a weak student is on their last warning - ignore it? Or perhaps use common sense and bring the student in to discuss.

No one is using any AI/plagiarism detector (and none are perfect) to be the final arbiter of anything. But where something is flagged, it needs addressing, not ignoring. That isn’t to say that detectors are all perfect, they’re not, and no one has said otherwise. But you can’t expect unis to just go ‘oh fuck it we won’t bother checking anything’. May as well give everyone a 1st and the email of an essay mill.

ForeverTheOptomist · 01/05/2026 21:41

Unissss · 01/05/2026 09:01

I wasn’t logged in to it as don’t have an account but surely anyone could have used ai multiple times not logged in for example or had multiple accounts

I'm sorry, but please, please accept that using AI or whatever crap is bad form. Really really bad form. Apart from anything else, you should be working on your construction of sentences and paragraphs. I started uni as a mature student, and there were countless people who were obliged to take a course in English language in the first year. What I'm getting here is that you don't believe your lack in written communication to be an issue. it is. And don't rely on AI in any way shape or form. You do not have a leg to stand on here, or cannot evoke any sympathy.

Gabby8 · 01/05/2026 21:42

Yes they can.

ForeverTheOptomist · 01/05/2026 21:46

Gabby8 · 01/05/2026 21:42

Yes they can.

Can what?

CreativeGreen · 02/05/2026 14:30

What an utterly bizarre turn this thread has taken.

8misskitty8 · 02/05/2026 14:48

Not sure what they use but colleges/uni it seems can tell if someone has potentially been using AI.
DD is studying web design and a couple of her classmates have had assessments flagged as possible AI and were checked further.
Are least one student has been removed from the course for it.

DonnaHadDee · 02/05/2026 15:09

It depends on the threshold for "reliably say". I work in the software development, and sometimes support the local Uni as an external examiner, so typically involves a sample review of MSc and PhD work. One typical check is a manual code review, or basically a scan through the student project repository. You can very easily tell the parts that are AI generated versus written by students. In these projects the use of AI is generally considered acceptable, but must be noted/flagged/credited by the student in some way.

Unissss · 05/05/2026 15:33

Had the meeting the update is they understood my explanation but did give me a grade penalty for poor academic conduct which is fair enough.

OP posts:
murasaki · 05/05/2026 15:48

That sounds fair. You were lucky.

Unissss · 05/05/2026 19:10

murasaki · 05/05/2026 15:48

That sounds fair. You were lucky.

I guess but I didn’t use it for writing it.

OP posts:
mids2019 · 06/05/2026 07:05

titchy · 01/05/2026 21:40

So if a lecturer ‘feels’ a student’s essay is AI written - what then? A feeling isn’t objective, so shouldn’t be used? A brilliant essay produced when a weak student is on their last warning - ignore it? Or perhaps use common sense and bring the student in to discuss.

No one is using any AI/plagiarism detector (and none are perfect) to be the final arbiter of anything. But where something is flagged, it needs addressing, not ignoring. That isn’t to say that detectors are all perfect, they’re not, and no one has said otherwise. But you can’t expect unis to just go ‘oh fuck it we won’t bother checking anything’. May as well give everyone a 1st and the email of an essay mill.

It's a large step in life to accuse anyone of cheating and as in other realms doesn't innocent until proven guilty apply? I think this is the problem as a lecturer may have a real sense of an AI supported submission but without rock solid evidence it is difficult to penalise anyone (and there may be legal ramifications to this).

the use of AI is a really grey one. I know of pupils asking to AI to check they have included all points within an essay and not change it per se so does this count as cheating? What if AI is used just to restructure existing work or suggest expansion? I think clever students/pupils will work this out.

to my mind AM a is brining us to a world where invigillated exams will be the only fair way of assessment sadly. Indeed I find in my daughter's school this world is soon coming.

Swipe left for the next trending thread