But the middle is the interesting part, and there is so much of it. In my school, we have everything, from students cheating well or badly to students who say they will never, ever use AI because of its environmental impact, and teachers using AI not, badly or well.
The hype, positive and negative, seems to be mostly a marketing strategy. There also is an unbelievable gap in ethics, which is dangerous.
Great article, Syd. I really like the metaphor of compression. We will find out at some point, perhaps in the not so distant future, whether this grand experiment on children will have net positives or negatives. From my limited experience, I am seeing signs that some kids understand the pitfalls of relying too much on AI tools. But one of my former students who is finishing his 2nd year at West Point described the Academy's approach and, perhaps unsurprisingly, they are permissive on many things while more rigid on others - he showed me some graphs he had made with AI and cogently explained where it's useful and where to be careful. It gives me hope that students will adapt and thrive. But the dangers are definitely there.
Stephen, thank you for reading — and for that West Point example. That is exactly the kind of nuance I think gets lost in most of the current conversation. A student who can show you the graphs AI helped him build AND explain where the tool is useful AND where to be careful is demonstrating the thing I am actually arguing for. He is not replacing his thinking with the tool. He is wielding the tool from inside his own understanding. That is the ownership layer. Once it is built, the extension is safe. Without it, the extension is the whole reasoning.
The compression metaphor came from watching the opposite pattern in my own classroom — students who can produce polished output in seconds but cannot explain the architecture underneath it. The cognitive work that used to happen during the time it took to reach an answer is the work that builds the architecture. When the time collapses, the architecture never forms. That is the compression I am worried about. It is not that students use AI. It is that some of them never do the building step that AI is supposed to extend.
Your read is right that some kids see the pitfalls and adapt. What I am watching for is whether we can figure out how to build the architecture in the ones who are not adapting on their own, before the window closes. That is the work that keeps me up at night. Thank you again for engaging — this means a lot coming from you.
New subscriber here. Thanks for recommending this. I just saw it earlier today (after reading your piece about it), and I’m very glad I did. I liked a lot about the way the material was presented (dialog between an expecting father & his pregnant wife, & also with AI industry leaders).
I’m not sure if the following are spoilers or not, so be aware: spoiler alert.
I am glad he gave time to the idea that you can’t just cordon off “good AI” from “bad AI,” and just enjoy the benefits of it. The knowledge that allows AI to detect cancer earlier than a human doctor, or even come up with a cure, is the same knowledge that will allow it to be used by a bad agent to construct biochemical weapons. Or even, that will allow it to develop its own, and use them against us humanoids.
I loved the first two sections of the movie (apocalyptic & optimistic visions of AI’s impact on humanity). I’m not so hot about the ending, which appeared to suggest that we need to get the government to impose an AI slowdown, until AI can be properly controlled. I’m not too sure I follow the logic here. If in fact an earlier statement/premise of the movie is true--that the good AI cannot be separated from the bad AI--then I'm not too sure how giving the federal government control over who does what with AI, and at what point in time, is the solution. And if we are supposed to orchestrate a grand, global coalition with the people who invaded Ukraine & gave us Tiananmen Square, I’m not sure that we can trust them to adhere to any limitations that we would be willing to embrace. So if it comes down to trusting Sam Altman or Xi Jinping, sorry but I’m going with OpenAI every day of the week.
OK I see my reasoning is probably a bit simplistic. Granted, I’ve written the above paragraph without subscribing to any stacks that promote the idea of a consensual slowdown. Perhaps I need to read a few. Any recommendations? If my line of reasoning above is in fact simplistic, I’d love to see what the counter is.
Regardless, I've decided to make the last unit in my modern history classes a discussion & presentation of AI. The documentary has given me some great starting points, so again thank you.
But the middle is the interesting part, and there is so much of it. In my school, we have everything, from students cheating well or badly to students who say they will never, ever use AI because of its environmental impact, and teachers using AI not, badly or well.
The hype, positive and negative, seems to be mostly a marketing strategy. There also is an unbelievable gap in ethics, which is dangerous.
Both Apple TV and Amazon Prime Video are currently accepting preorders for the movie.
Pity is has not been released internationally...
I can’t imagine it won’t be on a streaming service pretty soon.
Stephen — I've been following your work for a few months.
You named the missing middle perfectly. The documentary
shows what's at stake. What it doesn't show is what to do
about it at the student level before the bigger questions
even matter.
That's what I'm building. My article "The Integration Gap"
just ran in STEM Magazine (520K readers). My book is on
Amazon. Happy to send either if useful.
— Syd Malaxos
thinkinglabs.academy
I’d love to see it!
Here's the STEM article: https://stemmagazine.com/gAPRIL26/viewer/desktop
I’m on pages 22 - 25
And the book: https://a.co/d/0adownwx
Would love your take whenever you get a chance.
— Syd
Great article, Syd. I really like the metaphor of compression. We will find out at some point, perhaps in the not so distant future, whether this grand experiment on children will have net positives or negatives. From my limited experience, I am seeing signs that some kids understand the pitfalls of relying too much on AI tools. But one of my former students who is finishing his 2nd year at West Point described the Academy's approach and, perhaps unsurprisingly, they are permissive on many things while more rigid on others - he showed me some graphs he had made with AI and cogently explained where it's useful and where to be careful. It gives me hope that students will adapt and thrive. But the dangers are definitely there.
Stephen, thank you for reading — and for that West Point example. That is exactly the kind of nuance I think gets lost in most of the current conversation. A student who can show you the graphs AI helped him build AND explain where the tool is useful AND where to be careful is demonstrating the thing I am actually arguing for. He is not replacing his thinking with the tool. He is wielding the tool from inside his own understanding. That is the ownership layer. Once it is built, the extension is safe. Without it, the extension is the whole reasoning.
The compression metaphor came from watching the opposite pattern in my own classroom — students who can produce polished output in seconds but cannot explain the architecture underneath it. The cognitive work that used to happen during the time it took to reach an answer is the work that builds the architecture. When the time collapses, the architecture never forms. That is the compression I am worried about. It is not that students use AI. It is that some of them never do the building step that AI is supposed to extend.
Your read is right that some kids see the pitfalls and adapt. What I am watching for is whether we can figure out how to build the architecture in the ones who are not adapting on their own, before the window closes. That is the work that keeps me up at night. Thank you again for engaging — this means a lot coming from you.
— Syd
Thanks for the heads up, Stephen. I can't find it in the UK yet but am on high alert!
New subscriber here. Thanks for recommending this. I just saw it earlier today (after reading your piece about it), and I’m very glad I did. I liked a lot about the way the material was presented (dialog between an expecting father & his pregnant wife, & also with AI industry leaders).
I’m not sure if the following are spoilers or not, so be aware: spoiler alert.
I am glad he gave time to the idea that you can’t just cordon off “good AI” from “bad AI,” and just enjoy the benefits of it. The knowledge that allows AI to detect cancer earlier than a human doctor, or even come up with a cure, is the same knowledge that will allow it to be used by a bad agent to construct biochemical weapons. Or even, that will allow it to develop its own, and use them against us humanoids.
I loved the first two sections of the movie (apocalyptic & optimistic visions of AI’s impact on humanity). I’m not so hot about the ending, which appeared to suggest that we need to get the government to impose an AI slowdown, until AI can be properly controlled. I’m not too sure I follow the logic here. If in fact an earlier statement/premise of the movie is true--that the good AI cannot be separated from the bad AI--then I'm not too sure how giving the federal government control over who does what with AI, and at what point in time, is the solution. And if we are supposed to orchestrate a grand, global coalition with the people who invaded Ukraine & gave us Tiananmen Square, I’m not sure that we can trust them to adhere to any limitations that we would be willing to embrace. So if it comes down to trusting Sam Altman or Xi Jinping, sorry but I’m going with OpenAI every day of the week.
OK I see my reasoning is probably a bit simplistic. Granted, I’ve written the above paragraph without subscribing to any stacks that promote the idea of a consensual slowdown. Perhaps I need to read a few. Any recommendations? If my line of reasoning above is in fact simplistic, I’d love to see what the counter is.
Regardless, I've decided to make the last unit in my modern history classes a discussion & presentation of AI. The documentary has given me some great starting points, so again thank you.
Pro or con, the only bad arguments are no arguments.
Glad to hear of this.