First Student expelled for using AI

Post Reply New Topic

The University of Minnesota expelled a PhD grad student for allegedly using AI. Now that student, who denies the claim, is suing the school.




The higher education sector across the world is watching what the outcome of this investigation

I saw that. My personal experience with the disciplinary proceedings at college is that there's a lot of room for them to put their fingers on the scale to get the result they want. It certainly does look rather thin here, and it also looks like they didn't bother to actually contact the student before issuing the failing grade to see if he at least knows the material he was claiming on the exam.

(Side note, this is one of the reasons why I object so much to forcing colleges and universities to handle certain crimes in house, sometimes it makes sense, but often it doesn't)

That being said, whether he cheated or not, it's an absolutely damning indictment of the university to have such an important exam taken without proctoring. I remember when I was working on my masters that I either had to go into a professional proctoring location for some, or I had to be at home, in a room that was largely empty of anything that could possibly be used to cheat. I'd have to show the proctor all corners of the room with my webcam, point the webcam how they requested it and move it if the proctor later asked to see something. And, my cellphone had to be turned off and shown to the proctor that it was being turned off. There was also a program installed on my computer that was monitoring for processes that indicated barred software from being run until I clicked the final button to submit my test.

And, in this case, there appears to be no way of actually preventing somebody from using AI as it's not even proctored even though it's one of the last tests to be taken before being awarded the doctorate.

My understanding in the US is that many universities use some type of honour system for take home exams which is used currently in many Australian universities whereby students are told not to use AI. Of course we all know every student uses AI to draft take home exams/essays or learning system hosted quizzes.

For that reason I am interested in how this case pans out.

There are take home exams, but it's unusual for something this high stakes to be unsupervised. That may not have been the case around the time of the Pandemic and a bit afterwards, but the only times I ever had unsupervised exams was doing a 2 year degree, and the tests themselves were just a small part of the actual grade.

I definitely share the curiosity as to what happens here because the disciplinary process at a lot of American colleges and universities leaves a lot to be desired in terms of the normal protections that the accused get in court. They can definitely put their thumbs on the scale to get whatever result they like and failing to actually evaluate whether or not he had the knowledge that was expressed on the exam is may cause them some problems in court. Universities do have some latitude in terms of policies, but if the process wasn't handled properly, they may well be forced to settle and to let him back into the program.

Proctoring and eye tracking software is used for online tests/exams (a byproduct of COVID lockdown) but its not perfect, especially for large cohorts of students. I think a simple solution for this type of cheating is go back to paper and pencil exams on campus.

Didn't watch the video, because in my case videos aren't usually a good way to get information. So I found this:

https://gizmodo.com/minnesota-grad-stud ... 2000566900

Don't know what to conclude really, except that somebody is lying. My gut reaction is that it's the university that's corrupt, but I guess the student could be trying it on. It'd be interesting to see the transcripts of the court case, if the court is any good with due diligence. Some of those claims can be tested - e.g. "the professors edited the ChatGPT responses to make them read more like his answers."

I wonder how he got his lawsuit funded? I've known employees to sue their employers, but the union financed that.

The AI "art" that I've seen on Facebook lately has made my blood boil, but expulsion does seem like overreacting.

But then again, the USA in general is constantly either overreacting or underreacting to things, there's just no balance, everything has to be extreme.

Expulsion is a common punishment for academic dishonesty, when it's discovered.

It isn't just in poor taste like AI "artists", it's cheating and most schools have policies regarding academic dishonesty. The student here has almost certainly acknowledged that they understand they can be suspended or expelled for cheating so they received a punishment that they already knew was quite likely for the behaviour they engaged in.

I think both parties are at fault. He (the student) almost certainly used ChatGPT but the response from the university appears to be an overreaction. Worst should have been a zero for the assessment. I suspect (as Yoda would say) "More to this than meets the eye".

I suspect the lawyers will want tech logs of the student's submission in order to demonstrate the professor went in and edited the work. Perhaps the student was a nuisance and the UoM was looking for creative ways to expel him?

If you read the Gizmo article the student told the media he is using chat to write his legal case

^ This ^

Every syllabus I have ever received in college made it abundantly clear that if I engaged in academic dishonesty, which included representing someone else's work as my own, I will not receive credit for the work, I will receive an F in the class, and I will be suspended at the minimum, and possibly expelled. At a graduate level especially, they don't play around about it.

I have had to deal with this issue for a few semesters now. When I teach a course, I have a statement in my syllabus that prohibits the use of AI on the coursework. It lays out that someone that is caught using it will automatically receive a failing grade from my course and be reported to the student counsel about their actions. I know that some students use the functions of AI to help them build reports and papers. It becomes a chess game to me. To catch them, I always put in a technical question that AI would have a very hard time getting correct answers to (because it currently is giving wrong answers to that topic). Might it answer that way because someone put that incorrect answer in for a reason? Hmmm, yeah.

"To beat those cheating with AI, you must become sneaker than the AI that they are using to cheat with."

The University really isn't doing themselves any favors in how this is being handled. I do get that they have to identify cheaters and prevent them from succeeding in order to remain accredited, but it doesn't really look like they took adequate care to preserve his rights and the courts may well decide that regardless of the truth that his rights were violated.

This sort of thing is one of the reasons why I'm so opposed to having colleges deal with criminal offenses. It's one thing if a college wants to enter into an agreement to have a diversion program for minor infractions that could unnecessarily ruin lives, but handling felonies in house is a terrible idea as there aren't the same sorts of protections. Colleges are not in a particular position to render these decisions impartially and may not even have any sort of reasonable path to appeal.