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

[Unissss]

i personally don’t see how tbh

Aren’t they kicked out?

It’s likely that it’s left a bit of code on the page when you’ve copied it over.

It might have done. But don't you have university acceptable things for that, e.g. Grammarly, others are also available. It sounds like it reworded it rather than just do a spell check. Which is against most rules. If you can speak to the content if asked, you may be ok.

And remember if part of the assessment is exam based AI won’t help at all and a discrepancy between course submitted work and exam work will only raise more queries.

@Unissssyes potentially ( but of course there are multiple variables on how students use AI) the students are then sent down another process of academic integrity and what happens to them is decided by Chair of Board of examiners.

Don’t most document programs highlight spelling and grammar issues as you write?

If your problems are that bad, I'd suggest talking to the disability office re things that are acceptable to use.

You only mention checkers and flagging, OP. These are automated processes. Did a human assess your work? Lecturers have other ways of recognising AI. Are you being completely honest on here?

I would assume most likely because of the same algorithm of syntax, convoluted explanation and basic grammar. If you're experienced then you absolutely know if the student who produced such an essay really wrote it.

I'm so glad to be out of this sector of marking now because it's appalling to have students not attend sessions yet hand in phenomenal copied and pasted essays with barely a glimpse other than to add a few sentences. So hoping this isn't the case for medical studies, where a deep knowledge is fundamental.

Other 'art' subjects may be able to get away with it aa the outcome doesn't rely on saving lives/creating and maintaining safe engineeringmachinery/improving infrastructure with common sense advice.

Depends if they are lazy and just too incompetent to know how to use it.

It is trivial for someone who knows how to competently use them, to produce work that is 100% undetectable.

But the people using them to cheat are specifically doing so to put almost no work in, often have limited understanding of the topics anyway, and don't have any understanding of an llm or how to use it. So for most people cheating yes the university will be able to tell. If someone who knows what they are doing tried, not a chance.

Yes, but to varying degrees and it depends how lazy you've been in using it..if they suspect it they will call you for an interview to see if you can evidence understanding of what you've written and done the work. If you can't, you're cooked.

If you submit it without making any changes then probably yes. AI is often superficial and lacks real analysis; AI creates a word salad which often sounds impressive but on closer reading says very little; also if you sit in my lectures playing with your phone, don't prepare for seminars and look blank if I ask you a question then produce something worth a 2:1 I'm going to be suspicious. As @keepswimming38 says we have ways of spotting it.
I will call students in for a viva and they have to satisfy us that it's their work. My university has a three strikes and you're out policy. First year students are also treated more leniently but if they do it again especially in the second or third year it's treated far more seriously.
I've changed my assessments and on one of my courses students now do a presentation and on another I've introduced an exam. Presentations can be produced using AI but you've also got to be able to talk about the topic for 15 minutes and answer questions for another 5 minutes.

It's even funnier when they reference nothing on the reading list, and you can see their access to both online and in person use of the library......which is not related, if they've even used it....

Honestly, some of them think lecturers came down in the last shower.

I'm a lecturer. This would have been against the rules in my discipline at my institution. We don't allow any use of GenAI. Others have different rules. For us, this kind of usage - if a student admitted it - would be a mark deduction and not a fail.

As a PP said, we hold a misconduct hearing, ask the student to explain their work and show their research process. You can also share your internet and/or AI engine history to show how you were using it and that you also did research.

No. They think they can though and it results in some innocent people getting blamed for using it. I know that people with autism especially have a disproportionate level of being blamed for AI when it's their own writing style, something to do with how their brains work. I've seen issues with mature students who are used to writing in a professional manner too. It's a known thing, although those in authority like to deny it because it suits them not to have to think for themselves but just to scan people's work and have a computer decide.

I was coming to say this, really. Although work is anonymously marked, once marks are entered, you can see the names. Alarm bells ring with certain phrasing - and also, the ones who are too dim to remove the 'created by ChatGPT' tag.

And that is known, which is why in person conversations happen.

It's not just about writing style or scanning work.

Anyone who thinks that's all that's involved is going to get caught out.

That's not how it works. Computers don't decide. It's people who interview students and make the final decisions. It's much more than 'scanning'. It can actually be quite difficult to pursue cases because universities want to avoid complaints to the OfS.

I don't work in a uni but my role involves reading a high number of written submissions. There are very clear tells when AI has been used. For example, there has been a proliferation of references to certain laws and statutes, which we simply never saw before. I don't use any software, but over the course of the past year or so, it's become possible to make a pretty good guess at whether AI has been used. So yes, I think unis, via a combination of methods, will often correctly identify it.

You need to check your institution's policy. Where I work, we have three levels of outcome from an investigation:

• No further action
• Zero mark awarded, and student to resubmit (which means you can't get more than the basic pass mark for any resubmission)
• Full disciplinary, which can mean no opportunity to resubmit, and therefore failure of award, or requirement to withdraw.

(To be fair, this last option is rarely used, unless it's very severe, or a repeat offence.)

In my Uni, I undertake the secretariat service for a lot of disciplinaries and investigations. I would advise being as honest as you can, including if you did actually use AI. Honesty is usually treated more leniently than continuing a lie (not saying you are lying.)

I would add that speaking to your Student Union for help and support for any investigation. They have specially trained officers for this.
(Edited to include last paragraph.)

So they've accused you of it. There is a process they have to follow now. Have they explained to you what the next steps are?

Yes, I've done several. I reckon it was 10% NFA, 20% mark deduction, 70% submit with a capped mark at pass level. Saying that, none of them were final year.

I teach at a uni and the answer is no. AI checkers are unreliable, and while I can obviously tell something has been written by AI, I can't provide formal proof. So it's just my word against the student's.

If you have told them this then you ought to be given opportunity to supply them with the original unchecked version, so they can verify what you say and give appropriate weight to what you can claim as your own content.
Have you done this?