Frankly, I couldn’t care less. But Terry Mills cares and he asked me to ask.
I am suspicious of even pedestrian uses of AI, and I am deeply suspicious of deeper uses. Even if AI should be able to improve school education, I am far from convinced that it will. But the thing is obviously invading, and it will infect everything.
Whatever. Terry requested it and I’ve asked:
How is AI used at your school by students and teachers?
Go for it. I’ll check in after the footy.
Visited 41 times, 1 visit(s) today
Take ChatGPT as an example.
Last two years, we used it to verify the answers to some parts of General Maths exam paper.
We also used ChatGPT to generate some multiple choice questions for Methods, particularly Pseudocode questions.
It is a powerful tool to explore unfamiliar areas. But care should be taken, and further instructions must be fed to GPT to refine and polish the questions to a satisfactory and suitable level.
Thanks, Sentinel. These are exactly the kind of uses of CHatGPT etc that I had in mind (except for the pseudocode crap), and why I am so suspicious. What does “verify” mean? How much “refine and polish” is actually done? Without any accusations against your particular school, I simply do not trust AI in general to make questions better rather than simply making teachers lazier.
Writing and checking mathematics questions, even the basic bread and butter questions, to ensure that they are good and accurate is tricky and subtle. Most teachers (and most VCAA twats) suck at it. I see AI *maybe* improving the questions in sum, but I also have little doubt it will make the teachers worse.
Using ChatGPT for making tests is madness to me, especially because it makes frequent mathematical errors. Did you at least use one of the reasoning models like o3-mini?
AI is used by some of our students for cheating in exams.
Thanks, Franz. How do your students have access to AI during the exams?
One obvious horror consequence of AI is to kill valuable assessments where students can readily cheat with AI. The essay in Burkard’s The Nature and Beauty of Mathematics subject is a case in point. The essay, with students choosing their own (approved) topic survived as long as it did only because I was willing to grade 170-ish essays each year. (Burkard wouldn’t trust anybody else to do it.) I was willing to do this horrendous job only because I thought it was such a valuable opportunity for the students who took it seriously. This year, the essay is no more, precisely because of the many, many motherfuckers who cheated with AI last year.
Do you have a sense for how prevalent AI use is at school or university? I’ve got quite an ‘addictive’ personality so I avoid it, rightly or wrongly, because I’m worried I’d develop an unhealthy reliance on the tool and damage my ability to independently problem-solve. Maybe my fears are overblown here, but I’ve got the same concerns for primary and secondary school students – who perhaps now have a weaker* grasp on these ‘foundational’ skills?
*Not trying to be all “back in my day” about it! – this just seems like the natural consequence of having an unlimited tool that can do all your homework for you, though I have no idea of what actually happens in schools nowadays (esp from perspective of student use).
No general sense, but I think at university it is enough to be a disaster. With the maths essay last year, I suspected about a quarter of the two hundred students of having used AI, either a little or whole hog, but without working hard to try to confirm or reject my suspicions.
Wait the essay is no longer part of that unit? I remember when I was at Monash I took it in 2022 and the essay was my favourite assessment because it allowed me to research and write about the history of mathematics (rather briefly due to the word count) as I am very interested in it and find fascinating. Sad to see it gone. I genuinely despise AI in most contexts, but especially so in education… also thanks for marking the essays Marty that sounds terrible but hopefully you read some actually good ones sometimes
I just looked up your essay: you did a pretty good job.
Thanks for the comments, Emil. Yeah, the essay was a very good thing, which is why (and only) why I was willing to do it. In truth, the majority of essays were not great: too broad and formulaic and Wikipedia-ish. It was a draining task grading them. But each year there was a decent minority who recognised the opportunity and took it seriously. It didn’t mean their essay was necessarily strong, but there was the honest strong attempt (and it was marked accordingly). And, each year there were a few outstanding ones. And now it’s all dead.
I have no idea how they did it. I have kept a close eye on my black sheep during the next exams, and it worked.
On the loss of opportunity to do longer, more interesting, subtle assessments: this is genuinely sad in cases like Burkard’s course, but honestly I’d guess a net positive for high school.
‘Research tasks’ or ‘investigations’ in high school maths are uniformly awful. Even when kids aren’t cheating, there’s only so much deep thinking a year 8 can do about a single maths problem (or area) with the tools they have available. So invariably teachers end up scaffolding the bejesus out of them, and it just turns into a single, insanely-long worked example. Kids are graded on their persistence to get through it and/or their use of nice headings.
Next year, you’re looking over a kid’s previous grades and you see a glimmer of hope, “wow they got an A in trigonometry last yea- oh wait, that was the investigation. Never mind.”
Thanks, Alex, and a very good point. But I think it’s pure optimism to think that anything will kill such idiotic school investigations.
Hope springs eternal. Or at the very least we that we can limit the damage to the VCAA mandated slop.
Oh and on the main question: long-term very concerning etc. But in the short term I’m mostly finding it to be a really useful wedge to make my colleagues rethink their idiotic “Kids can look up facts on Google. What they really need is [some complex skill that the teacher has no viable way to impart]” stance.
Are they really rethinking? If they were stupid enough to ever believe that, what might they believe now?
The problem with teachers isn’t their dogged belief in bad ideas. It’s their inability to hit a sweet spot behind cynicism and credulity.
The attachment to these ideas is incredibly thin, but they’re invoked out of habit. 9 times out of 10 when they trot out some half-understood line about “building deep conceptual understanding”, the teacher they’re talking to nods along and they get the positive feedback they’re looking for. Eventually it becomes an instinct.
It’s a troublingly polite profession.
I, for one, am doing my best to alter that.
In high school, is it common for students use AI to “explain” assignments to them? It seems to be common in university with first years. The hopeful part of me really wants to believe that they simply don’t know this is cheating, and that if they just knew it was not allowed they would stop. Is that at all possible? Is it allowed in high school?
Thank you everyone for your responses. I found them to be quite interesting.
At my (secondary) school some teachers use this and say it’s great. I haven’t looked into it.
https://chalkie.ai/en-au
Thanks, Drumlin. Looks great: