Earlier in the week, I found myself at a conference: EduTECH24. Yeah, me neither. Being there felt pretty much like a Jew finding themself at the Vatican.
How I came to be there was pretty weird but it doesn’t much matter; I was there. It meant I could say hi to a couple cherished colleagues, and glare at a couple prominent assholes. And it meant I could observe the selling of edutech. There was a lot of selling.
Apart from the selling and the selling and the selling, there were lectures. Which was more selling. But there was also a maths ed stream. Which was held in a closet out on the back porch. The maths ed lectures were to varying degrees interesting or predictable or nuts in the expected manner. One lecture, by a very young teacher, stood out for its clarity and intelligence and humanness. I think I likely disagreed with the message but the teacher was as impressive as I’d seen for a long time. But maths ed and teaching, by teachers, was of course just a conference sideshow. The showcasing of impressive young teachers was the least of priorities.
Back at EduTECH Proper, the main drawcard was the Khan Academy guy, with the warm-up act being the Grok Academy guy. (Alan Commie-Hunter Finkel was also prominent, for God knows what reason.) Mr. Grok was appalling: ingratiating, unfunny, unclever and entirely unconvincing. He was pure salesman and a very bad one. Mr. Khan, however, was different. It took a real effort to dislike him.
I didn’t dislike him. Sal Khan spoke beautifully, with modesty and heart and humour and intelligence. It was probably a stump speech that Khan had given a thousand times before, but it didn’t mean that the speech was insincere or unmeaningful. It seemed to be neither. Khan told the story of the humble, accidental birth and then huge growth of Khan Academy. Then Khan got on to his main purpose, to sell Khanmigo, his new collaboration with OpenAI, the Chat GPT guys.
It seems clear that Khan values teaching, by teachers, and that he has the noblest of intentions. As Khan sees it, there is invariably one teacher and N kids. So, Khanmigo is a cheap and easy means for the students to have at least some human-ish feedback, while waiting for the human feedback. Khan talked about the details, of the teacher control over everything, and so forth. It is an alluring vision. Until you wake up and realise that it’s all nuts.
Khan is seeking to solve the problem of insufficient mathematics teachers and insufficient good mathematics teachers, but that is not the problem. Well, yes, it is a problem: way too many teachers suck and those that do not suck are asked to do way way too much. But it is not the problem. The paucity and poor quality of mathematics teachers is merely a symptom of the fundamental problem: that society has lost the meaning of education and consequently places no proper value on it.
Australia is rich. Of course Australia could afford to train and to employ sufficiently many properly good teachers if only the country woke up to what that means and gave a damn. But it does not. We do not. So, always the brilliant suggestion is some brainless technical “solution” to a cultural problem, which inevitably makes matters worse. This does not mean that there can be no role for Khanmigo and the other technosaviours, for now or always. There are clear advantages to automation and to pseudo-human assistance, when human-human assistance is in short supply, obviously now and probably always. But that is not the way this stuff is sold and that is not the way that it is bought. Technology is not assisting teaching: technology is supplanting teaching.
Edutech24 was, to say the least, a strange experience. And a thoroughly depressing experience. Seeing close-up the massive amounts of money being poured in to further pervert education was utterly heartbreaking. It is time to read some Neil Postman, or to watch some Neil Postman, grab a large vodka and go to bed. Maybe we’ll wake up tomorrow and it will all have been a bad dream.
What is the solution? Beyond each of us, in our constrained sub spaces, doing the best we can? Genuine question…I just haven’t taken to genAI, (and I use plenty of ML based math when needed in technical research (never in education), and when I am crafting it for exact reasons in a particular research problem, and not using it like a black box). GenAI is math transactional-ised for the economic machine – genAI cannot, by construction and definition, teach math in a way so that the kids can understand this, beyond just the math: the fact that the human motivation behind teaching and doing math is fundamentally orthogonal to genAI teaching math, who takes the ethical / moral responsibility for the consequences (including mr khan, whose videos I do like)
Thanks, SS. In regard to what we can do, I think not much more than you suggest. We can keep our eyes firmly fixed on reality and, locally, we can fight the good fight. We’ll lose the war but at least we can win some battles, and at least we can know we battled for the right team.
AI is very interesting and I have little interest in it. I am fortunate that I have very little low-thought writing to do. But without doubt it’ll be revolutionary. And, if nothing else, millions of people are going to be screwed over.
As for immediate positive or “positive” uses of AI, I know there is tons. Burkard uses it constantly, to write first drafts of administrative crap. And I know teachers use AI to write (hopefully) drafts of lesson plans and the like. I know teachers are swamped and I completely understand why teachers much value such uses of AI. I am very very very far from convinced that AI is remotely as good at this as teachers wish to believe.
“like a Jew finding themself at the Vatican”. Well it could be worse it could be a Jew in a mosque”. I have seen several interviews with Sal, and he makes a very good impression.
Yeah, I figured it was better to stick with tradition and go with the Catholics.
Safe 🙂
One proposal. Worry more about having a few really good maths students than fixing everything.
It is a much more achievable goal. It ensures it will still be possible to know the difference between good and bad math education in a generation. It is easier to expand from this than a diluted approach.
I offer what the folks at aops.com achieved in the US as a model.
The downside is that approach relies mainly on parents funding it so is inequitable. However, that could be countered by others supporting it. Many could do more to help by funding access to aops than by spending their own time directly.
It is also a means to demonstrate what is possible with access to high quality teachers.
It’s not ideal but a pragmatic way to make a difference.
Stan
i think this is a terrible suggestion.
Suspect you’re spot on Marty. My understanding is that AI is good at some niches. The decision-makers mistake this for a panacea (in the 90s, the solution to every problem in the public service was a website, then an app, then blockchain etc etc). And the decision-makers rarely have any real understanding so a simple solution to a complex problem – bingo! And the tech bros think they can solve any problem …..
And how to get society to value education properly rather than as a commodity is a good question.
And if some solutions actually lie in the past, then who could admit that? I almost choked on my cornflakes reading a recent article in the Guardian that said that explicit teaching was invented in the 60s!
Hey Marty,
I went to this too, but not in the paid version so I missed out on all those Maths sessions you took in! I agree that Sal Khan was the most impressive speaker of the day and his story around leveraging technology to somewhat produce the effect of tutors is good. As someone working with children who do not have much access to effective teaching this is certainly of interest to me. Not sure yet whether Khanmigo will really do this.
There seemed to be a bit of consensus that we should slow down on bringing in AI for students but begin to train teachers in it. This seems right but hard to imagine how this is to be done. I’m actually returning to study to get a sense of AI myself. I am particularly interested in using AI for diagnostic assessments and in assigning suitable fluency work.
Thanks, YT. I’ll note that “working with children who do not have much access to effective teaching” is somewhat understating your admirable work, but I’ll leave that be.
Of course you, and Khan, are right, that there is no value in being pure about this. If the best that is possible involves a large dose of techno then that is what you do. However there are large howevers, and no one with power, or much of anyone, appears to be listening to these howevers.