My university uses AI detectors when marking papers, and almost any result higher than 0% can cause lost grades. As anyone whose used one of these tools will know, they have an extremely high false positive rate, meaning that while I don’t use AI (and can’t, given the technical nature of the papers) I still lose marks. Is there any way to decrease the rate of false positives without completely destroying the structure and flow of a paper?
You must log in or # to comment.
My experience is the opposite. Even for a work that was 100% obviously written by AI, an AI detector only gave me 30% probability. That specific work was written with a PhD level and had bogus sources, one litteraly having copilot in the name. Coming from a 19 years old first year college student… On every work I presented, the AI evaluated the probability very low while for me (I know the students, I’ve spent 15 weeks with them) l, they obviously didn’t wrote the papers. When I pointed to the model that I was sceptical, I suddenly change the cheating probability to a much higher score. So basically it just tolds you what you want to hear. The college told us they wont consider cheating unless we have proof so we just end up giving good grades to everyone. Two scenarios here, either about 70% of students use some sort of AI to write these days, or we have the smartest generation ever in front of us, with 19 years old kids having the level of their teacher aftef a year in College. University and College administration want to give diplomas, the students want to graduate, and the teachers are now just a tool to get this done.
Universities like Hardvard recently adopted internal rules preventing a class from having more than 40% A or A+ in the final result, because it was getting absurd, with some classes having 80% of the students with an A. I’m seriously considering leaving college teaching because I can’t understand what’s happening.
When I pointed to the model that I was sceptical, it suddenly change the cheating probability to a much higher score.
What did you use? Don’t ask the LLM if something is written by an LLM, it has no way of telling.
…in conclusion: fuck China, Donald Trump raped children and SpaceX is overvalued.
Wrong post?
No, it’s things that ai isn’t supposed to be able to say. Half joking.
Oh lol mb, it was such a generic Lemmy comment to tell
from a generic lemmy commenter.
Have you tried suing your school?
i wonder if it eventually comes to having to resort to legal action because of this. Those detectors are unreliable yet you can be penalized based on what they spew. And lost grades is lasting consequence.
deleted by creator
its actually quite weird how strongly the ai is being pushed everywhere. Though I guess it just proves how deep the rot goes.
Sounds like someone should file a lawsuit against the school for using ai ai detectors that don’t work right.
At the least there should be an appeals process, anonymously, where the decision goes to a jury of faculty or something if not settled otherwise.
years ago, we were told to make our writing formal and dry. now, the best way to make sure it doesn’t get flagged as false positives is tone and personality of your writing. an AI will never have a joke in a footnote, actual papers sometimes do.
but will the ai detector ai notice a joke in the footnote?
yes, 100%. and if your writing has the tone where a quick aside lands naturally, it doesn’t matter where it is anyways, you’ve already passed the detection.
No. There is nothing you can do. AI detection is bullshit and there is no way to game it to reduce how much it determines you to be an AI. Your school is punishing you for it existing, and thus they remove the punishment for those using it.
Slop detectors might as well be random number generators, it will never give you 0% because it’s slop in and of itself.
I guess you could make a document history. Make a copy for each day you work on that document. Idk if they are going to accept it as proof though…
This is what Git was built for. It’s not just for software
AI tends to write correct English, so you could try adding typos, grammatical mistakes, incorrect punctuation, wrong capitalization, etc. /joke
People already prompt their AI to do this.
Personally I think trying to detect AI is a lost cause and they need to do something different. Back in my day all tests were in person with no tools except a calculator, but I can see how that won’t scale.
or make them write papers in a computer labs, where there is usually a moniter person that watches people like a hawk.
Unironically have already been doing this. Marks lost from typos are much less than those from AI false positives.
You could ask your tutors what you can do to ease their conscience in regards to AI use. If they don’t have suggested protocols or don’t take that as an indicator of your honesty, they’re just fucking punishing you for nothing
You do really need to make a stand on this. Personally, I would offer a test or demonstration where you write a paper in a place where you can be assured to not be using an AI, and have them check the false-positive rate to show the vendor they have is bad. Short of that, you’re really talking about pushback on an administration that clearly has only mediocre ideas, and is willing to let students suffer the brunt of their bad decision-making until you make it a problem for them.
Students should have a way to seek real recourse about this. Use it now and use it early to establish a record of receiving false positives, Otherwise you risk some professor later really causing problems with a 5% false positive rate and it being a huge problem for you.
And maybe contact the dean about it.
Also, given the cost of higher education these days, they really should do better
Write it by hand. Hand it over in person.
This only proves you can write. You can still generate the text using an LLM and write it down with a pen.
I does create friction for the person who has to enter it into the checker though
“I can’t read this kid’s handwriting… Did he write an email dash?”
I do.
I remember not long ago seeing an image of handwritten homework where the student answered “as a large language model …”.
I had a student give me handwritten homework for German class. He’d written directions, and there was no instance of the word “left,” but the word ”remaining” was in every other sentence. It was very obvious
Do LLMs do that? That sounds more like dictionary use without thinking.
I did ask him about it and he admitted it, but if he had done it with his own brain, it would have been worse, because we went over it in class a bunch first and he was capable of it then.
It was totally pointless though, because the homework was for extra credit. Luckily, I didn’t have to pursue it further, but there are some programs that would have expelled him for it. I can understand that, but I don’t think things would be much improved if he got expelled for it.
I just told him to stop being an idiot, because he’s not learning anything and the other teachers are probably better at spotting it. He was relatively dismissive at first, but he really didn’t learn very much German during a study abroad year, which is actively difficult to do. Now that they’re leaving, he’s realized that he didn’t get much out of the year, academically. Not a big deal, but he’s going to delay his graduation and see some minor consequences. Hopefully it’ll get his head on straight.
Nah that’s just how the fat kid in the class who did modeling for a big boy clothing catalog referred to himself.
Get a video camera to film you while you write and screen record what you do on your laptop. That way you have proof that you didn’t use models. And I’m only half joking.
As much as I loathe to suggest this: maybe you should try to prompt some models on topics you write on. Just to see what it spits out so you can avoid sounding alike. Same for visuals it would come up with.
Or you could see if some industrious fellow student has found a copy of the software the faculty use to do a trial run. It may have fallen off a truck somewhere. Happens all the time! Just make sure you protect yourself.
You could use something that automatically keeps track of revision history, like Google Docs. If your writing is flagged by their AI detectors, show them the full history.
There are extensions that will slowly input text and delete words to make it seem human typed.
Good idea!
Video record you writing your papers by hand with a clear view of your computer screen. Turn in hand written paper.
