Another fantastic piece of thinking and writing; thank you, thank you, thank you. We are genuinely lucky to have you.
Doesn’t AI (LLM) fit hand-in-glove into the industrialised, corporate context education? In other words, the mode / model of many industrial educational systems (measurement against State-imposed arbitrary standards) makes something like chatGPT especially attractive; the ability to create a consistent product is one of the goals of some of these systems.
Your mention of Freire speaks directly to what education and learning /should be/. God, I love his ideas. Yet the context we find ourselves in (born from Horace Mann, Ellwood Cubberley, and Frederick Taylor’s influences) seems to resist the notion of humanist education the UN (and Freire) espouses. In addition, the standards-and-accountability era (A-Nation-at-Risk → NCLB → PISA) helped to create the policy space AI is now entering.
As a high school classroom teacher, I am pushing back on the use of LLM to assess student work, and students. But it is such a powerful tool and so easy for students to substitute thinking for a product. As long as learning is assessed as a product, tools like chatGPT will continue to dominate and replace our industrial model of learning.
I know I haven’t addressed that school systems vary widely; some already privilege process, metacognition, or community problem-solving. Also, teachers, departments, and exam boards can—and in many locales already do—re-design tasks so that mere text output is insufficient for credit (e.g., live performance, iterative design reviews, reflective viva voce). I also haven’t discussed LLM used as a Socratic interlocutor or to surface counter-narratives which can support that aim when assessment rewards the dialogue.
But my main point is that LLM’s fit quite nicely into “learning as product”. How do we change that myopic view of learning?
That's lovely, Bill, thank you. I'm lucky to have such engaged readers!
Two quick thoughts - the socratic interlocutor systems I've looked at are built on top of the basic models - out of the box I don't find them at all useable for this purpose. I think that's (one of) the holy grails for the edtech industry as dinstinct from the AI industry more generally and I'm planning a podcast or two on how that effort is going and what the costs and benefits might be.
I think in general LLMs fit into 'learning as productivity' which is very similar to your idea. And I hope my very broad category of 'repurposing' gives due respect to the creative work we are all doing - with our students - to bring forward other modes of production besides text. But whatever the mode of output, almost every task requires preparation that can be done using LLMs. And LLMs can produce in other modes - students are vibe coding answers to design problems, for example. So text isn't really the problem, much as the industry wants to tell educators it's all our fault for valuing reading and writing as modes of thinking (so let's just get over literacy and see how democracy works without it shall we?) The problem is productivity as the goal, which is antithetical to deep learning and self-development.
A comprehensive and timely response to the continued excesses of A'I'. Thank you, Helen. As someone (perhaps you?) said in Dublin a few weeks ago, the problem with A'I', (just as with most LMS (Learning Management Systems) is that they are not 'learning' systems, but rather 'compliance' systems, worthy of the most extreme fascist systems in history - and that goes right back to the infamous (?) Graham 'crackers' (sic) of three centuries ago. Time for a change, but how?
I am trying, in my own small way, to implement open, collaborative, revenue-sharing practices on Metalabel / Substack, on Shakespeare's dramatic work (of all things!), starting with Macbeth (for what it's worth.)
LoKAIT: Inviting Educators to Shape the AI-Powered Teaching Tool for Remote Communities
Exciting progress from the LoKAIT project!
Our rugged, solar-powered AI education device is now moving into hardware development — thanks to engineering partnerships with Soluware.de (Schulz Group) and production support from EmbedIT Solutions. The device is designed to work fully offline, bringing STEM, literacy, and life skills education to children in remote, underserved, and crisis-affected regions.
But hardware alone is not enough.
We are inviting educators and learning designers to help shape LoKAIT’s pedagogical model — especially in the training of lightweight AI models that support local learning, critical thinking, and cultural relevance.
LoKAIT is not a teacher replacement.
It’s a teacher’s tool — designed to support educators in areas where resources are limited, schools may not exist, and connectivity is not an option. With your expertise, we can guide its learning logic to reflect what truly works in the classroom — even when the classroom is a tent, a tree, or a village square.
If you’re an educator, curriculum developer, or edtech innovator passionate about educational equity, we invite you to contribute your voice to this global mission.
Another fantastic piece of thinking and writing; thank you, thank you, thank you. We are genuinely lucky to have you.
Doesn’t AI (LLM) fit hand-in-glove into the industrialised, corporate context education? In other words, the mode / model of many industrial educational systems (measurement against State-imposed arbitrary standards) makes something like chatGPT especially attractive; the ability to create a consistent product is one of the goals of some of these systems.
Your mention of Freire speaks directly to what education and learning /should be/. God, I love his ideas. Yet the context we find ourselves in (born from Horace Mann, Ellwood Cubberley, and Frederick Taylor’s influences) seems to resist the notion of humanist education the UN (and Freire) espouses. In addition, the standards-and-accountability era (A-Nation-at-Risk → NCLB → PISA) helped to create the policy space AI is now entering.
As a high school classroom teacher, I am pushing back on the use of LLM to assess student work, and students. But it is such a powerful tool and so easy for students to substitute thinking for a product. As long as learning is assessed as a product, tools like chatGPT will continue to dominate and replace our industrial model of learning.
I know I haven’t addressed that school systems vary widely; some already privilege process, metacognition, or community problem-solving. Also, teachers, departments, and exam boards can—and in many locales already do—re-design tasks so that mere text output is insufficient for credit (e.g., live performance, iterative design reviews, reflective viva voce). I also haven’t discussed LLM used as a Socratic interlocutor or to surface counter-narratives which can support that aim when assessment rewards the dialogue.
But my main point is that LLM’s fit quite nicely into “learning as product”. How do we change that myopic view of learning?
That's lovely, Bill, thank you. I'm lucky to have such engaged readers!
Two quick thoughts - the socratic interlocutor systems I've looked at are built on top of the basic models - out of the box I don't find them at all useable for this purpose. I think that's (one of) the holy grails for the edtech industry as dinstinct from the AI industry more generally and I'm planning a podcast or two on how that effort is going and what the costs and benefits might be.
I think in general LLMs fit into 'learning as productivity' which is very similar to your idea. And I hope my very broad category of 'repurposing' gives due respect to the creative work we are all doing - with our students - to bring forward other modes of production besides text. But whatever the mode of output, almost every task requires preparation that can be done using LLMs. And LLMs can produce in other modes - students are vibe coding answers to design problems, for example. So text isn't really the problem, much as the industry wants to tell educators it's all our fault for valuing reading and writing as modes of thinking (so let's just get over literacy and see how democracy works without it shall we?) The problem is productivity as the goal, which is antithetical to deep learning and self-development.
Now I want “the problem is productivity as the goal” on a t-shirt!
and allied to consumption, which leads, almost inexorably, to compliance, and a desert of creativity and agency.
Thank you, Helen, for this optimistic, bold and practical manifesto for human education.
Thank you John, I plan to make more of 'repurpose, rebuild, refuse' in other posts.
A comprehensive and timely response to the continued excesses of A'I'. Thank you, Helen. As someone (perhaps you?) said in Dublin a few weeks ago, the problem with A'I', (just as with most LMS (Learning Management Systems) is that they are not 'learning' systems, but rather 'compliance' systems, worthy of the most extreme fascist systems in history - and that goes right back to the infamous (?) Graham 'crackers' (sic) of three centuries ago. Time for a change, but how?
I am trying, in my own small way, to implement open, collaborative, revenue-sharing practices on Metalabel / Substack, on Shakespeare's dramatic work (of all things!), starting with Macbeth (for what it's worth.)
LoKAIT: Inviting Educators to Shape the AI-Powered Teaching Tool for Remote Communities
Exciting progress from the LoKAIT project!
Our rugged, solar-powered AI education device is now moving into hardware development — thanks to engineering partnerships with Soluware.de (Schulz Group) and production support from EmbedIT Solutions. The device is designed to work fully offline, bringing STEM, literacy, and life skills education to children in remote, underserved, and crisis-affected regions.
But hardware alone is not enough.
We are inviting educators and learning designers to help shape LoKAIT’s pedagogical model — especially in the training of lightweight AI models that support local learning, critical thinking, and cultural relevance.
LoKAIT is not a teacher replacement.
It’s a teacher’s tool — designed to support educators in areas where resources are limited, schools may not exist, and connectivity is not an option. With your expertise, we can guide its learning logic to reflect what truly works in the classroom — even when the classroom is a tent, a tree, or a village square.
If you’re an educator, curriculum developer, or edtech innovator passionate about educational equity, we invite you to contribute your voice to this global mission.
Join us. Shape the AI. Teach the world.
Contact: lokait2025@gmail.com
More info: https://open.substack.com/pub/abrahamjoy/p/can-a-machine-ask-a-good-question
#LoKAIT #EdTech #AIForGood #Team4Tech #DigitalEquity #OfflineLearning #HumanitarianAI
Superb!