ND daughter wrongly accused of using AI

[EmeraldDreams73]

Hi, dd21 is ND and in her 3rd year of uni. She's been accused of using AI to write her most recent assignment. She's not very tech-savvy at all and is absolutely adamant she hasn't used it and wouldn't know how. I believe her, she never does anything against the rules and she can't lie, there is no way she's capable of just brazening it out, she would 100% crack if she'd done it. I've read the submission and I'm no expert, but it just sounds to me like her, but trying to sound good.

She's had a meeting with the module lead, took her handwritten notes and the synopsis she'd written when requested earlier in the process and talked about it. Also highlighted words in the finished submission that she remembered finding through googling "better word for...".

She knows the content inside out, but always uses Word instead of Google docs for drafts - she once lost a big chunk of work at A level when she accidentally highlighted the wrong bit on Google docs by mistake and refused to use it from then on. She saves and renames files, but deletes old ones once she's got a final version - she's anxious about submitting the wrong one. She took her (v old) laptop in to the meeting and they obvs couldn't see digital revisions (she didn't know you had to turn on the Word tracking thing on that version, I think they showed her how so that's good going forward).

From what I can gather the meeting was inconclusive bc the lecturer must have seen she was being truthful, but still felt it was too different from previous papers (hadn't used grammarly etc, they asked - again, she wouldn't lie). She's been told that their head of xyz needs to review it and it's 50/50 what will happen next. Then heard nothing more for 24 hours which I'm sure is normal as they're all busy.

She's beside herself with worry. She has a processing disorder which means she takes forever to get things firmly in her mind, but then is fine. She's always had extra time for assessments/exams and does at uni too.

She can't remember what she said in the meeting and is terrified that they'll "think she's stupid" and kick her out. Even making her rewrite it would set her back hugely regarding time for current assessments, and the whole thing has just trashed her confidence. She was really proud of that one as she'd spent even longer than usual "going over every sentence and finding better words to use and I put loads of commas in".

I've told her that if they ask for another meeting, to say that she'd like me with her. I really want to ensure they know the kind of person she is, as well as the kind of learner, and the damage this is doing. She's already stressed about workload, but was pleased with how she was managing.

Anyone had experience of this kind of thing? What was the outcome? Her focus has always been to keep saving things to avoid losing work but (stupidly, I suppose) it's never occurred to her or me that she'd need to prove she's not cheating.

Actually a pretty standard identifier of an assignment that has been done by AI as a one off is that it differs completely in style, punctuation and syntax etc from the same person's previous assignments.
So your daughter could refer them to her other assignments, and they ought to be able to see substantial clues that it's the same person. There are hundreds of giveaways, for example use of commas, overuse of a favourite word, length of paragraphs, paragraph openings etc.
She needs to point out these similarities.

If she cannot demonstrate this then I too would be suspicious I'm afraid.

@ShyMaryEllenah. I thought it also checked for AI as well. I was worried about it when I recently handed in my dissertation. (Hadn’t used AI)

One of the main markers for AI use is that the grammar and spelling are perfect. It has a very smooth slick style that is quite difficult to replicate by students. All AI detectors are salute nonsense and it is embarrassing that universities are using them. They’re also seems to be a lot of misinformation with people thinking that if you have put your work into AI, Turnitin will pick up on it. It will not.

They’re also seems to be a lot of misinformation with people thinking that if you have put your work into AI, Turnitin will pick up on it. It will not.

Absolutely. In fact a very low score on Turnitin is often a red flag - because the work hasn't quoted from anything that exists.

https://blogs.ncl.ac.uk/sin/2025/08/05/the-unfairness-of-ai-flagged-academic-misconduct-investigations-in-uk-universities/

No. I probably shouldn't say this, but 'checking' for AI is all but impossible, as each cheat is written separately so there is no record of it to check against. All that can be found are 'indications', which are turns of phrase and so on. There are often the usual clues such as all US sources, dated citations and journals that aren't i the university library - but they are the sort of things that make markers run checks rather than a form of proof.

OP said they did check against her previous assignments and it was markedly different.

I mean, there are some dead giveaways - hallucinated references, em dashes, certain sentence constructions, the word ‘interplay’ etc. But any online detection doesn’t work. I have heard lecturers say that if students put their work into chat gpt then it is effectively ‘on the internet’, and that’s how turnitin flags it. Nonsense.

Thanks everyone, I appreciate the replies. Of course they won't believe my (or her) word on her integrity, but it's a small uni and they know her well. I've sent them an email just saying please bear in mind that she's ND, has always had extra time for exams etc and her thinking is v black and white, her sole focus is not appearing stupid but she'd never think about being thought a cheat. I said that she had spent a particularly long time on this one (she used all of Word's grammar/spelling suggestions etc). She also told me a few weeks ago "I put SO many commas in this one!" bc she was actively trying to use more to help with clarity.

If she has to rewrite it, she'll do that but it's such a kick in the teeth when she works for hours on something to improve then they don't believe it's her!

I went to uni an age ago and they had plagiarism checkers then (had to be under 10%). I think the ones they have now must be even better tbh.

TBH, you don't need to be tech-savvy to use ChatGPT, and it's not complicated. It's just like asking Google a question, and then saying yes, yes, yes to its offers of follow-ups. The key thing is that it generally tells you what you want to know.

Anyone using ChatGPT needs to click on the links it provides to verify the information, and that the direction it is taking you is where you want to go.

I don't really see a big ethical problem within higher education. When I did my degree in the 80s, I had to spend hours in the library looking for papers. AI just speeds up this process. As long as you are not having AI write your final essay, and you are the one that is summing up the information, I think this is an appropriate skill for now and the future.

The modern ones are very good, but a decent essay has a lot of citations and quotes, and the bibliography can run to more than 10%.

Again, though - using AI to write an essay is not plagiarism (which is basically copying). It is getting the essay written for you. Turnitin cannot (as yet) detect AI. I would like to think that something will be developed soon that can, as if not there will be a drive to move back to essays written in exam conditions, with all the angst that that entails.

They aren’t great. For example, a nursing essay will use (NMC, 2018) numerous times so that’s going to increase the similarity with other nurse students’ assignments. The title will often be the same as all the other assignments, so that increases similarity. If the student has a cover sheet to say this is the work of a student with specific learning difficulties, that increases similarity. A teacher trainee will use Education Endowment Foundation and Department for Education along with all the other teacher trainees. Psychology students will use APA or BPS and Turnitin will say “1% of this essay is the same as an essay from the University of Norwich” or whatever.

We never had those issues and my essays were covered with NMC references. Never went above 10%.

Generally up to 25% is OK now.

Turnitin uses an AI writing detection system that analyses writing patterns to estimate the likelihood that text was generated by large language models such as ChatGPT or Gemini. It also claims to detect Quillbot and Goblin Tools.

My autistic ds has had similar accusations.

It’s because of his use of words.

So - for example - he calls the sea the ocean and because he’d use The Ocean it’s assumed it’s AI writing and he’s not checked and changed it from (American) AI.

I hope it gets sorted soon.

I know😀. It can also find Grammarly likelihood. But it is an indicative result, not proof, unlike Turnitin, or even Google Advanced search, which will find the exact source of plagiarism.

Did she use copilot to help her as that is an AI based program within word that can help. For an ND person, this could cause confusion as they wouldn’t necessarily think they had used AI as the resource is available in word.

That's a thought! Her Word version is old so it might not have it (?) but that might make sense if it's on there and she sees it as the same as google.

I do not think a parent emailing a university about an issue concerning a 21 year old third year student is helpful to the student at all.

I agree, and the odds are (rightly) that nobody would act on it anyway, but I don't see anything wrong with a mother advising her daughter.

Fair enough, but the ND context was the reason and is v relevant imo. She was in complete meltdown when I spoke to her, couldn't remember if she'd explained her thought process or not. Perhaps they're well aware anyway and I didn't need to email, perhaps they're not and they still won't read it or care. She's much calmer now but very depressed :(

They will probably care, but staff are not able to engage in discussions with others about adult students. In very limited circumstances, where the student has given consent it may be possible, but not otherwise, which is fair enough, IMO. It can be hard if you think a student is suffering, but university staff are not in loco parentis, and it would not be appropriate for them to speak about them to their families.

Many people of student age are in work - would it be appropriate for parents to contact HR departments? It's a similar concept.

She has now heard from student support saying the matter has been closed. We've just had a chat about it being a GOOD thing to sound like herself! Maybe this week will give her the confidence to spend less obsessive time on redrafting in future, and just focus on getting the content down! It'll save her so much time too, which can only be a good thing at this stage of her studies.

I also had a lovely email from student support saying they couldn't discuss anything with me (obvs) but thanking me for the context re ND. Doesn't matter if the powers that be never saw it, but it was kind of student support to reply.

When dd2 is at uni (which she will be next year), I will of course stick to my previous hands off approach, she is NT and doesn't have difficulties. But given dd1 has struggled for her entire education, I'm glad I at least tried to advocate for her. Thank you to everyone who took the time to reply.