UCT scraps AI cheating detection

mylesillidge

mylesillidge

Journalist
Joined
Jul 29, 2021
Messages
4,268
Reaction score
4,562
Oh lovely. Vibe coders, doctors and engineers incoming...
Patient will be bleeding out and the doctor can't do anything because the "system is down"...
 
mylesillidge said:
One of South Africa's biggest universities drops AI cheating detection

The University of Cape Town (UCT) has scrapped the use of artificial intelligence (AI) detection software as it shifts towards ethical AI use.

Speaking to Newzroom Afrika, director of UCT's Centre for Innovation in Learning and Teaching, Sukaina Walji, explained that AI detection tools aren't reliable enough to use confidently.
Click to expand...
Lowering the bar even more...

Now you don't even have to bother getting 30% on your own...
 
Most (not all) degrees aren't worth the paper they're written on these days as information is so easily and readily available.
 
VegOtter said:
Oh lovely. Vibe coders, doctors and engineers incoming...
Patient will be bleeding out and the doctor can't do anything because the "system is down"...
Click to expand...
That AI detection is just barely better than guessing.
 
However, he said this isn’t the only way in which AI affects tertiary education. He warned that many students don’t even attempt to engage with prescribed course material.


As a result, he said, many students are graduating with degrees, having never been thoroughly assessed or taught critical thinking.
Click to expand...
Ok but - isn't it the role of the university to develop methods by which these things are taught & tested? Sit the student down in a room and hammer them with questions.
 
Regarding the self diagnosis thing using AI, when it comes to those GP's that just send you through the conveyor belt of expectations without giving a f**k about what might actually be wrong with you and just stuff antibiotics in your hand while sending you on your way, I tend to trust AI more than I do some of them.

Saved myself over R65K when an "expert" insisted I get pins for a broken collar bone, did my own research by comparing my plates to some of the published research on the topic and decided to ignore what he said, just went in every two weeks for some plates to make sure the bones are healing as expected.
Doctor told me after 6 weeks I made the right choice.
 
Last edited:
saor said:
Ok but - isn't it the role of the university to develop methods by which these things are taught & tested? Sit the student down in a room and hammer them with questions.
Click to expand...
That's what exams are for. But it's never resulted in students having a real grasp of the material.
 
RaptorSA said:
Regarding the self diagnosis thing using AI, when it comes to those GP's that just send you through the conveyor belt of expectations without giving a f**k about what might actually be wrong with you and just stuff antibiotics in your hand while sending you on your way, I tend to trust AI more than I do some of them.

Saved myself over R65K when an "expert" insisted I get pins for a broken collar bone, did my own research by comparing my plates to some of the published research on the topic and decided to ignore what he said, just went in every two weeks for some plates to make sure the bones are healing as expected.
Doctor told me after 6 weeks I made the right choice.
Click to expand...
My wife was only diagnosed with Endrometrias, after I read a thread on it in these forums. The symptoms matched, we went to a professor who specialised in it and it was confirmed, after an operation.

Previously doctors just basically downplayed her concerns.
 
Segg said:
Most (not all) degrees aren't worth the paper they're written on these days as information is so easily and readily available.
Click to expand...
Degrees are about a lot more than availability of information though. If by information, you mean tools willing processes that information and package up essays, projects, and quizzes, for you, I would agree.
 
Coming soon to UCT:

the-adventures-of-dr-mcninja-omnibus.jpg
 
saor said:
Ok but - isn't it the role of the university to develop methods by which these things are taught & tested? Sit the student down in a room and hammer them with questions.
Click to expand...
There is so much credit given for tutorials, practicals, and projects, that most course results are dominated by these. E.g., a student may get 70% of their marks from these, and then 30% of their marks from the exam. They may in theory get 100% for their practicals, and 0% for their exams, and leave with a 70% average, despite never having done any of the work themselves.

Some courses have minimum exam marks for passes (e.g., usually 40% when present), but how many have this, and how well it is enforced is very dubious. In my opinion, exam marks should count much more (possibly everything in today's world).

Swa said:
That's what exams are for. But it's never resulted in students having a real grasp of the material.
Click to expand...
Agreed. I think that tutorials and practicals, should still be given, and still be a requirement (DPR) to write the exams, but I don't think they should contribute to the course marks anymore. For larger projects (e.g., final year projects, honours projects, etc.), I believe that these should be included, because using AI as part of these would simply be more realistic. Ideally these projects should be complex enough that AI can assist, but not build the significant components of their projects.
 
It is unreliable because it is detecting original and genuine work as AI-generated. I created some graphics the other day, manual work, and an AI detection tool validated it as AI. Copywriting tools also now have their inaccuracies.

Source material is now in doubt.

To be more on topic. UCT students still need to do examinations.
 
RaptorSA said:
Regarding the self diagnosis thing using AI, when it comes to those GP's that just send you through the conveyor belt of expectations without giving a f**k about what might actually be wrong with you and just stuff antibiotics in your hand while sending you on your way, I tend to trust AI more than I do some of them.

Saved myself over R65K when an "expert" insisted I get pins for a broken collar bone, did my own research by comparing my plates to some of the published research on the topic and decided to ignore what he said, just went in every two weeks for some plates to make sure the bones are healing as expected.
Doctor told me after 6 weeks I made the right choice.
Click to expand...

rh1 said:
My wife was only diagnosed with Endrometrias, after I read a thread on it in these forums. The symptoms matched, we went to a professor who specialised in it and it was confirmed, after an operation.

Previously doctors just basically downplayed her concerns.
Click to expand...

If your issue isn't stock standard, most doctors have never, and would never be able to spend as much time on your (or your partner, or kids') particular issue(s) as you can. Between forums, online research, and LLM's, I think it is very reasonable that once your doctor/specialist has hit a wall, that you (as someone capable of understanding and contextualizing the data sources) are more likely to find a root cause or treatment by your own agency, then by continuously begging for tidbits of hope, asking them "What next?", after each failure.

I have had several successes at this myself for sick family members.
 
You must log in or register to reply here.
Back
Top
Sign up to the MyBroadband newsletter
X