Professor thinks I’m dishonest because her AI “tool” flagged my assignment as AI generated, which it isn’t…
Wednesday, 12 Mar 2025
Exclusive insights, data, and analysis for financial market experts.
Search
Have an existing account?
Sign In
Those ai detection softwares are the polygraph of academia.
This happened to me with a multi page final term paper (college) that I had indeed spent days writing myself. None of it was plagiarized or AI-generated, and the “AI Detection” program rated it like 88% AI-generated or something like that. Those things are trash and shouldn’t be used.
Lol we had to use turnitin in the 2010s in public school. It had this “similarity score” that was supposed to theoretically detect plagiarism.
In reality all it detected was that you mirrored the question in your answer the same way 17000 other students did. It was so trash, that I didn’t know of a singular teacher that actually gave that number any credence whatsoever. So it essentially was a massive waste of everyone’s time.
I saw a lot about turnitin during covid. It would seem in the 10 years or so since I had used it, it hadn’t gotten any better, and I doubt “AI checkers” are any better. Also when you consider the problem itself of developing an AI to detect AI, you begin to understand what a fools errand it is. Unless we mandate that AI includes identifiable watermarks of some sort I doubt it’s very solvable.
The irony of professors being overreliant on technology to counteract overreliance on technology… I get there’s not really a better way to do it but it just feels wrong to act like chancing unjust punishment is better than chancing unearned success
Have you used software capable of showing the history of your writing? From rough draft to final edit? Seems that now a days one almost has to show every step they took in writing the paper themselves (with timestamps if possible) for it to be truly believed.
I’m the stubborn kind of jackass that would say
Sorry not gonna do it , I didn’t use AI
Fail me and call me a liar or let it go
The purpose of AI training on human written text is to simulate human written text. Thus AI recognizing human written text as likely AI, it’s simply detecting it’s source material.
Garbage in, garbage out.
I’m very glad I am not in academia during these years…dealing with plagiarism was bad enough!
Find a paper your professor has written, run it though an AI detection tool, and then send them the results. I’m very sure it’ll be flagged as AI generated.
Unfortunately we are at the point where a lot of people’s writing is so bad that coherent, well written sentences get flagged as AI. Best of luck to you OP, hopefully you’re able to prove that you wrote it yourself. Can’t imagine how frustrating that must be
Submit the declaration of independence, the AI detectors are *really* sure its 98% ai generated.
Honestly at this point it seriously seems like it’s more a tool to teach kids to advocate for themselves when they havnt done wrong than to actually catch cheaters.
The cheaters know they cheated and won’t argue, and the ones they didn’t will get mad and take it up with the prof.
These tools are useless and are more likely to give a false positive than a true result. Most readily admit to that or can easily be proven to be so by running pre-ai or the professor’s own content through them.
I’m a lawyer and occasionally do defence of academic offence cases. It’s appalling to see how many academics use these, often resulting in frivolous blanket allegations of dozens of students. Imo profs should be sanctioned for using this shit, academic disciplinary tribunals are often one-sided enough without the faculty gaslighting students with false “evidence”.
No, I’m not rewriting a damn thing. That’s obscene. AI can’t write intelligently. It’s usually very obvious when an AI generates a text body. It should be very obvious at a glance whether something is artificially generated.
As lecturer that has to constantly deal with students cheating with AI, the eye test is more reliable than the software. I can tell when a student has used AI to write entire sections of their work just by the style of the writing, so I doubt she read it herself. When they have edited the AI writing, it is more difficult though. Ask your teacher to read it and forget about the software score and if they are 100% sure that it is AI written. Tell them your thought process, and that you don’t want to rewrite it because you don’t want them to think you are a cheat, because you’re not.
Setup word to track revision history. With AI “detection,” tools, this is probably the best way to defend yourself against false positives.
You need to submit a complaint to the head of your department or the Dean of students. It’s not very academic for your professor to act on an assumption, and frankly unprofessional of her not to try to talk to you in person first. This will likely continue happening until you do.
Plug the declaration of independence into an AI detector and send her the results
Screen record yourself writing I’m future, feels like the only way to prove it was hand-type
A second camera behind would also work
Yeew. What subject does this professor teach?
“from what I have excitedly heard” makes me cringe. Does “more than capable” need hyphens?
I get the feeling that “warning” is AI generated. So I guess time to fire the professor because I’m sure they sent that out more than their 3 strike system allows!
Hopefully there is a way to dispute this or outright sue them. Good luck.
Many of these tools are trash & even if not most users don’t understand how to use them or what the results mean.
Nobody should be giving warning unless they can first exhaust all other explanations, to do less is lazy.
Recently my son got such a warning because his work rated high on the school’s tool & was completed in a single shot with little editing. Thing is he uses assistive speech to text tools & a grammer / spelling checker that does not integrate well with the school’s software so he has to cut & paste.
At this point they may as well ask us to screen record the writing process.
~~Please write a 2 page, single-space report on xyz~~ “Please send your 29GB, 3-hour uncompressed raw recording of your 2 page assignment.”
Google docs has a revision history thst shows when and how content was added and in blocks too. So if you take a break for a second BOOM history note, you step away for a nap or whatever BOOM note. Show them those notes as it can show line by line and even character by character in some cases as well as your fuck ups and revisions
Run the professors email through an AI detection site as well and just send the results back to them
Rules for thee but not for me
What service were they using?
This won’t change until a university is sued – or the developer of the software is sued.
These programs infuriate me to no end. The entire concept is the most braindead thing ever. You have AI that is getting better at sounding “human” while AI detectors claim to be getting better at detecting AI. In another 5-10 years these things will just be flagging everything anyway
This happened to me over the previous semester. A 10-page paper in the U.S. Constitution and Bill of Rights. Said the whole thing was AI and plagiarized. I had to fight it and get an appeal. Can’t stand tech and cheaters making it harder.
Even the basic plagiarism checkers are flawed. If you submit your paper to one to make sure you didn’t accidentally plagiarise, then send in a revision, it will get an increased plagiarism score because it has seen your document before. Your professor then runs your final document through, and you get a 75% plagiarised score.