A professor used white font to hide instructions to throw off students using an AI chatbot to cheat on their midterm.
If they developed a way to catch students taking photo of their screen there would be 3x more caught.
32 out of 35 is pretty much all of them.
The really stupid thing is that the students are the ones that are cheating themselves. Higher education is not cheap and cheating your way through it leaves you coming out the other end stupider and less qualified to take on the world as an adult.
Shit is supposed to be hard. That’s called learning.
Now here is the thing. I’m terrible at writing. Plus I have ADHD. English has never been my strong suit. I’ll start a thought, get sidetracked, start another thought, leave out words, etc. I’m an engineering, math, science kinda guy. Case in point, back in the day (like 40 years ago) my SATs were very skewed. My math score was like 780 but the verbal was 200 points lower. If AI had existed back then, I would have written the paper or whatever outline of my own ideas, and then had the tool make it coherent. LLMs are good at language. That’s what they really are. They are not smart, they don’t have original ideas. They are word guessers. I’d treat it like a spell/grammar checker on acid. Having the LLM ( I really don’t like calling it AI because it really isn’t AGI… at least not yet) do the entire assignment is just short cutting the entire process of what the university is there for.
Just my 2¢
weird opinion
The midterm, by the way, was about the Industrial Revolution.
I get it, you spent a fortune on this class so you could have a job that will allow you to maybe retire someday. You don’t even give a shit about this class or the material, you don’t care about history and you just needed the credit. It has literally nothing to do with any of the jobs you might get in the future, if you learn nothing it won’t make a difference. If you study and try your best you still might not actually get a good grade. The temptation to cheat would be hard to resist.
All AI has done is reveal a deeper problem with college - students aren’t there to learn, they’re there to be credentialed.
Yes and no.
Yes, degrees are credentials and sadly many people just want those instead of having any inherent interest.
But no, because those credentials are not empty words, but are supposed to mean something. You might not need to know about the industrial revolution in your job (and looking at the way our world develops right now, I would say industrial revolution and its changes in society are very relevant right now), but the credential for it is supposed to show that you are able to learn about a topic, read relevant literature and then apply that knowledge in analysing things. THAT is the goal of going to college. And most jobs (edit: that really require a certain college degree) need exactly that kind of skill. Which students don’t learn if they let a probability machine do the “thinking” for them.
Ok, but just to play daivels advocate: If you can’t tell they are cheating without tricks, can’t they do that learn+read+apply thingie using AI just as well???
That’s the question, isn’t it?
I think for some things it will work out, but as we are seeing with programming: the problem isn’t apparent in surface level, but in real understanding. And if a certain task need an understanding to apply knowledge, LLM can’t help with that.
Shit, this also reminds me of Manna, a short story about AI that starts out by micromanaging workers. In a way that is what LLMs are doing here.
Interesting story, thanks for sharing. Let’s hope something like the Australia Project is possible in our future and not only Manna.
Pro-tip: If you’re young, immigrate to Taiwan and get citizenship, then you’ll eventually be annexed by China during the collapse lol.
