I can’t be the only one who reads both this blog and some of the deeply-in-the-heart-of-AI ones like Astral Codex Ten.
What many of the true AI believers in SF/Silicon Valley think about education is actually scarier even than “screens help kids learn” or “jobs are all screen based anyway”. For many of them, they think that AI is a superhuman form of life that will soon (like, before 2030) take over most thinking and once AI itself designs better robots, most real world jobs too.
Half of them then think that humanity is cooked and the robots will kill us, and half think we’ll be kept as kind of like primates in a zoo. No need to learn to think or write or anything because what’ll be the point? The artificial superintelligence will do all of it better and faster anyway.
Even the tech-industry folks who don’t go this far still believe that little armies of AI agents will do all our thinking for us and take all actions, write your emails, etc. Also common is a belief that the thing to do is to become part of the capital-owning class as soon as possible since AI will supercharge wealth polarization even more than it already is.
All of which is to say — these people do not have our children’s best interests at heart. It all needs to be out of schools as soon as possible. More broadly: Our thinking is what makes us human. If we outsource that, what’s the point? (No wonder they don’t care about schools. They’ve given up on the future of the species, many of them.)
In developing knowledge of and experience with the use of AI, such as for research purposes, of how much of a machine it is, and a thinker it is not. In the end, a human should have to sign off on anything to propose it to be true.
I have surely seen the AI support a position, and then support the exact opposite position, and this can branch in multiple directions. I am reminded of what is needed to write a good research paper in 1950, or whenever, of that which has not changed. The best question I've ever asked an AI is, "What does the opposition say?" That is also the best question I have asked prior to the use of any AI. Tackling the topic from as many different angles as possible, and hunting down greater numbers of scenarios for context, with or without the machine, contributes the same quality to the result. Yes, the machine can wade through mountains of material in seconds, but only I can truly perform the high-level function, which will be to say that whatever is being said is ultimately true or false.
When you pull back the curtain of the screen, you are still left with a learning culture that has learned how not to learn. The learning culture is, and has always been, play deprived. That's why the temperament of young people is conditioned to gravitate towards the shiny things, the low-hanging fruit in the lives of everyone these days. The learning culture is a vacuum, a terrain of depleted soil where conversation could thrive if play propelled the classroom ecosystem the way it propels human intelligence. Play as a force of nature needs to be fully present in the hands and hearts of all the participants. An infrastructure composed of play -based communication is called for to make play central to the learning culture.
Hi Sophie. I worked in educational publishing for several decades. I came to it from the world of journalism. It was an absolutely wonderful experience that met a tragic ending: our company, an old and venerated one, was slowly destroyed by private equity. But that's another story. What surprised me was how fad-prone the field of education is. In the beginning, I would go to teachers conventions every year and every year the leaders in the field would be pushing another yet another new approach to teaching, often something involving the latest technologies. I grew very cynical pretty quickly and stopped going to the conventions.
Vert true. This fad cycle in education can be traced back more than a century (and is documented well in chapter 3 of “Education and Technology” by Neil Selwyn) with the rise and fall of educational film, radio, and TV. There are striking similarities between these technologies (which we would consider low-tech now) and today’s technologies. The problem is that none of them address root problems in education and rarely focus on the appropriate context to actually deliver education.,
Corporations have influenced schools for a very long time. By the late 1800s and early 1900s, wealthy industrialists began investing heavily in education.
Not because they funded everyday public schools directly through taxes—that remained primarily the responsibility of local and state governments—but because they shaped what schools taught and how they were organized. Major philanthropists included: Andrew Carnegie, John D. Rockefeller and Henry Ford. Their foundations financed:
Teacher training
Universities
School libraries
Vocational education
Educational research
Curriculum development
The Factory Model of Schooling
If we're still pretending that these philanthropists were simply trying to give back, we're being incredibly naive. Each of these men built their fortunes in industry, grooming immigrant adults and their entire families, how to behave in factory environment. This is when schools became organized around:
Bells
Fixed schedules
Grade levels
Standardized curriculum
Age grouping
Efficiency
Standardized testing
Large classrooms
The industrial model of education was never designed to cultivate independent, critical thinkers. It emerged during a period of rapid industrialization, when the economy demanded large numbers of workers who could reliably follow procedures, arrive on time, respect hierarchy, and perform standardized tasks. Schools increasingly mirrored factories: bells signaled transitions, students were grouped by age, learning was standardized, and success was measured by compliance and the ability to produce the "right" answers.
Critical thinking, questioning authority, and challenging accepted ideas were not the primary objectives of this system. In many ways, they were inconvenient. Imagine a factory floor filled with workers who constantly questioned every instruction from an authoritarian supervisor. Efficiency—the defining goal of industrial production—would grind to a halt. The educational system evolved to support the needs of that economy, rewarding obedience, conformity, and consistency far more often than curiosity, creativity, or independent thought.
That doesn't mean every educator wanted to suppress independent thinking. Many exceptional teachers have always encouraged students to ask questions, explore ideas, and think for themselves. But the structure of the system itself was built around standardization rather than intellectual exploration and capacity. The result is a legacy that still influences education today, even as the skills most needed in the modern world—critical thinking, adaptability, collaboration, and innovation—are the very qualities that standardized systems have often struggled to develop.
Many historians describe this as the "factory model." It mirrored industrial workplaces because that was considered the most efficient organizational system available. One well-known example was the Gary Plan, developed by William Wirt in the early 1900s.
Students rotated through:
Academic classrooms
Workshops
Gymnasiums
Auditoriums
This allowed schools to educate more students using the same buildings and reflected efficiency principles popular in industry.
Frederick Taylor and Scientific Management. The work of Frederick Winslow Taylor greatly influenced schools. Taylor developed "scientific management," which emphasized:
Standardization
Measurement
Efficiency
Specialized roles
Productivity
School administrators adopted many of these management practices, although classroom teachers often adapted or resisted them.
Were Schools Designed to Create Factory Workers? This is one of the most debated claims and as someone who has spent a fair amount of time in schools, I struggle with the idea that this ideas is even debatable at this point.
Schools were not created solely as factories for producing obedient workers. Industrial efficiency was one influence among several.
Who Actually Paid for Public Schools? The vast majority of funding has historically come from:
Local property taxes
State governments
Federal government (a much larger role after the mid-20th century)--also heavily influenced by corporations and their lobbyists.
Corporate "philanthropy" influenced educational policy and innovation and it's doing this today, pervasively.
Many features of today's schools reflect their 19th- and early 20th-century origins:
Age-based grades
School bells
Standardized testing
Fixed schedules
Uniform curricula
Seat-time requirements
If we hope to transform public education, we'll need to persuade the organizations with the greatest influence and resources—including major corporations—that investing in schools designed to cultivate critical thinking, creativity, and human potential is also an investment in their own future.
Tomorrow's workforce will not be defined by how well people follow instructions. Increasingly, success will depend on the ability to think critically, solve novel problems, collaborate across disciplines, and innovate in ways machines cannot. These are the very capacities that flourish through the arts, music, design, theater, and other creative pursuits that have too often been treated as expendable rather than essential.
The same is true for physical well-being. An active body supports an active mind. Research continues to demonstrate the connection between movement, exercise, attention, emotional regulation, learning, and mental health. Recess should not be viewed as wasted instructional time, and physical education should no longer be treated as an afterthought. They are foundational to helping children develop into healthy, engaged, resilient learners.
The irony is that the very skills and capacities many employers now say they desperately need—creativity, adaptability, emotional intelligence, collaboration, resilience, and sound judgment—are often the first casualties when schools become narrowly focused on standardized testing and measurable outputs. If corporations truly want a future workforce capable of navigating complexity and driving innovation, they have a vested interest in supporting an educational system that develops whole human beings, not simply efficient workers.
Your closing point is the part worth sitting with longest, though. Persuading corporations to fund a different kind of education is a real, necessary fight, slow and structural, and it may take generations to actually shift. But a person doesn't have to wait for that fight to be won before they stop running the obedience pattern the system installed in them. That work can start now, in one person, regardless of what the system around them does or doesn't become.
Thanks for sharing. I take a different approach to evolution and change that doesn't start with a fight. The future is generated by inspiring vision--rather than fighting with what is. In terms of school transformation, that work is well under way in cities all across the US and in other places. Moving it to another level by including the stakeholders, (corporations), that benefit from more humans who have the skills and desire to think critically, is where we are now.
I don't think there's actual disagreement here either. Inspiring vision, bringing corporations and other stakeholders into a shared future, works at the level of the system. What I'm pointing at works at the level of the person, right now, inside whatever system currently exists.
Those aren't competing paths to the same destination. They're two different things happening at once. The system can be slowly reshaped by vision and inclusion, and a person can stop running an old pattern in themselves without waiting for that reshaping to finish. One doesn't have to fight the other for either to be true.
I appreciate what you're saying. From the top of a big hierarchy, uniformity serves.
In an example from the article (LA), it took a citizen group to influence the school board in the right direction. More primary schools can be influenced more locally by parents/citizens. This is doable in an age where the general population appears to have no effect at a federal level. That would take national citizen organization - I think more likely, state-wide organization is doable.
Yes, I agree. Only, most parents these days, with both working, at least one job, maybe more, don’t have the space to dive into the PTA, or whatever it’s called these days, to advocate for these sense making changes. I’ve spent a fair amount of time working in environments with undeserved children and heard the comment that “these parents just don’t care” far too many times to stomach.
In a quick search, I did find this and will share that I agree with this 100%—When I bought my first laptop (early adopter over 30 years ago), I envisioned a time when more people would be working from home than going to the office. Who knew Covid was coming and the surge of remote workers. As someone who came out of the womb organizing social communities, I pondered that this would then require a heightened sense of community because human beings need socialization and if people were not engaged at work, they would need to create other spaces. Then came co-working spaces and an abundance of coffee shops that somewhat fills this need, but I haven’t seen either, and I’ve belonged to two co-working spaces that I enjoyed, but they’re really missing the mark on intentional community as they center only on workspace and ways to grow a business…. and coffee shops, at least where I live, miss the mark altogether and as a business consultant, that’s a strange dynamic to witness. Every shop and co-working space would profit more by taking this on…. “Robert Putnam's theory centers on social capital—the idea that social networks, norms of reciprocity, and trust hold communities together and make democratic societies function. He famously argued that a decline in civic engagement and community involvement leads to poorer governance, increased polarization, and widespread social isolation.”
Thanks for your note. I’m not familiar with Putnam’s work, but I do know that a lot of scholars and others have written about the loss of community and I have tons of my own theories as well. I like the thought of there being local groups that are helpful. Say more.
The main article on my 'stack is in 3 parts, and part 1 is an introduction to Putnam. Since that writing, from the AI there is also the idea of a CAPED group: Community Anchor for Participatory Ecosystem Development:
Asset Mapping & Shared Resource Management
Incubator for New Mutual Aid & Civic Groups
Community Concerns-to-Opportunities Pipeline
Rotating Leadership & Skill-Building
(Along with the idea in my article)
That's general thinking for now. As I say, I think we are getting into a position to think of making a list of health-related reasons for local groups.
Sophie, this is exactly right, and there's a mechanism underneath it worth naming. Learning isn't information transfer, it's the slow cultivation of a rich, personally-threaded memory substrate: the "fibrous journey" you describe. Handwriting, paper, boredom, a teacher's eyes on you: these all force the child to do the threading themselves, latent to active, effortfully.
Unfortunately, I currently teach math online mostly for the past 6 years since covid. Autocomplete or a chatbot tutor doesn't help the child weave their own mind together, it substitutes an external one. The engagement mechanics you name (streaks, rewards) aren't a design flaw bolted onto good pedagogy, they're what happens when a system optimizes for attention capture instead of memory formation. Because I teach online, I rely mostly on grading notebooks, with lots of practice with pencil and paper. This is my workaround. And project-based learning.
AI fingerprints are all over your comment. "there's a mechanism underneath it worth naming. Learning isn't information transfer, it's the slow cultivation of a rich, personally-threaded memory substrate: the "fibrous journey" you describe." There is irony in replying to a article about tech taking over schools...with obvious AI output.
We must restore critical thinking, actual memorization, handwriting, and collaboration. This is what I do with my students. "Fibrous": as in build our brains again through play and work and interactions with others. I long deeply for a day when AI chatbots are gone. I use them to edit my work. I make a lot of mistakes. But I'll be so happy when they are gone. There's a reason I still do not own a cell phone. There's a reason I'm reading this substack.
And why did big tech participate in undermining the education system to abandon the common sense that children cannot change their sex? Might attempts to change kids' exterior appearance with drugs and surgeries be a really bad idea that spreads like a contagion harming families and society on a mass scale?
Sophie — thank you for this essay. The engagement-metric slot-machine critique and the "personalized learning as depersonalized learning" quote from John Warner are two of the sharpest lines in the field this year, and I'll be citing both of you repeatedly.
For readers who stayed with the piece to the end: the products Sophie is describing — Chromebook autocomplete assistants, chatbot tutors on canonical content, engagement dashboards — are what I'd call controller architectures. AI deployed above the student, prescribing from outside. Complex Adaptive Systems theory of AI tutoring predicts they will fail exactly the way Sophie describes them failing: they fall into what my paper under review at BJET calls surface-compliance and mutual-capitulation attractors. Sophie's diagnosis is right; CAS names the mechanism.
There is a different architectural category — AI as constituent agent participating in a course's local rules, on a high-friction doer-effect substrate. My FIN 345 finance students at Ensign College compute CAMEL ratios by hand from raw 10-K filings; the AI persona works alongside them, not above them. It sustains the crawling phase Sophie rightly wants to protect.
Sophie's critique is required reading. The architectural alternative she may not have encountered is real, and it works.
Your research on these types of AI sounds very interesting! What AI tool do your students use that works alongside them? Have you seen any other examples of AI as a “constituent agent”?
Brian — good questions, and I owe you honest answers on both.
The tool is Crucibl, the methodology and multi-agent architecture I've been building since 2024. It's currently deployed in FIN 345 (Financial Institutions) and FIN 485 (finance capstone). Four AI personas across the scaffold-fading arc — Parker (Full), Avery (Shared), Rowan (Light), Riley (faculty-facing) — each holding rules at the turn level: anti-fold discipline, completion-claim verification, audit-trail visibility. The RGAS architecture is specified in three papers currently under peer review. Public methodology overview at crucibl.tech.
On other examples: honestly, the explicit constituent-agent framing is not widely established yet, but there are adjacent efforts worth knowing about. Bastani et al. 2024 documented in a Turkey GPT-4 RCT that unstructured GPT-4 access produced 17% worse independent performance while a scaffolded configuration matched or beat the control group — the difference was architecture, not capability. Kestin et al. 2025 achieved Bloom two-sigma outcomes in a Harvard physics RCT with a structured tutor that resisted surface-compliance by design. The broader intelligent-tutoring-systems tradition (Kulik & Fletcher's 2016 meta-analysis) captures the pre-LLM version of the same architectural idea.
What's different in the Crucibl case is that the four failure modes are named explicitly and the rule-level countermeasures are architected in from the start. Whether that's the right way to frame the space is what peer review will tell me.
What a smart woman! She also happens to be a prominent member of the British Royal Family, yet unlike another actress member of that family who sells jam with her title, she never mentions it. That's because she is not an insecure arrivist.
Thank you Sophie for your powerful voice in this fight. EdTech and AI are also destroying the relationships that kids need with their teachers. “School connectedness,” which is driven by kids’ bonds with their teachers, not only promotes learning success but it also prevents kids from becoming depressed and suicidal. The last thing this increasingly depressed and suicidal generation of kids need is a screen that gets in the way of a real-world connection with a teacher. The first thing my kids talked about coming home from school was a story their teacher had told them. Such connections are what inspired them to do the hard work of learning. Every child deserves the opportunity to connect with a caring teacher, not an unfeeling laptop.
I think this is written by a knowledgeable person. I want to throw this angle in there. We have a rash of cases where there are emergency situations, and someone should dial 9/11, and people bust out there phones and film it instead. I see it looks like they view themselves as videographers, doing something useful, but also with later fallout, including either feeling shame in the following days for not actively participating, or by becoming more morally hardened.
In this sense, there is the question of what do screens do to the way we behave? We propose "cognitive offloading" in the case of information tools, which means we're not to remember anything. Plato didn't like the idea of writing for this reason, in a strongly oral time. I am also reminded of the US military's "information cockpit", which uses 7 screens that give context to the information being displayed, and increases memory retention by 55%, they say. I am reminded of reading a book, and remembering that something was halfway down the page on the left-hand side, maybe halfway through the book. That gives it a kind of context.
But screens are so much more of a way to fundamentally change our relationship with our world.
I believe, without checking that Plato merely recorded that Socrates advised against teaching writing.
He himself was literate. Socrates left no actual writings.
As i noted this way back in 9th grade i may be in error,but my memory ( moving often, i retain no Platonic materials) is that your comment concerning Plato is in error.
I asked the Google AI, and it said what's below. Perhaps this is also interesting in context of this article.
Plato argued that writing is bad for society because it weakens human memory, creates the illusion of wisdom, and destroys the interactive nature of true philosophical debate. He detailed these concerns in his dialogue Phaedrus through the myth of the Egyptian god Theuth (the inventor of writing) and King Thamus. [1, 2, 3, 4, 5]
Plato’s main critiques of writing include:
• It weakens memory: Plato argued that relying on written texts makes people forgetful because they stop exercising their natural memory. He viewed writing as a crutch rather than a tool for true learning. [1, 2]
• It lacks interactive dialogue: In a conversation, you can ask questions, clarify misunderstandings, and debate ideas (a process known as the dialectic). Plato believed written words are static; they cannot defend themselves if misunderstood or adapt to the reader's needs. [1, 2, 3, 4, 5]
• It creates the illusion of wisdom: Written text allows information to spread easily, giving readers the illusion that they are knowledgeable without actually going through the mental effort of understanding the concepts. [1, 2]
• It falls into the wrong hands: Writing cannot distinguish between who should and should not receive the information, broadcasting complex philosophical ideas indiscriminately to people who may not be ready for them. [1, 2]
Instead of writing, Plato championed living, spoken dialogue as the only way to achieve true philosophical enlightenment and deep understanding. [1, 2]
Children are NOT small adults. They are in the process of becoming!!! Some say until 25, others until 32. You listen to them not for wisdom but for freshness and surprise. Skipping the process parts leads to challenge conditions for that person
See the literature on dendritic pruning as well as that on PfC connectivity and myelinization.
Yes, 25 or so is the supposed final connective maturation.
The role of astroglia and other glia is being refined in neuroscience, and th full story of differential maturation , while outlined, is an important area of research.
We mature in quite distinct stages. See that well-known literature to help yourself become less tentative in comments.
Peripubertal adolescence , when salient attention largely shifts to peers is a VALUABLE stage in which adult-imposed cultural norms is and SHOULD be HIGHLY CRITICIZED by adolescents.
You will find that rational self interest is highest in that 13/4- 25 stage. After that submissiveness to obviously aberrant social structures and cultures, overcome actual intelligence.
I truncate this comment with the proviso that nearly all adults are, sometimes HIGHLY, deficient in understand what should be basic knowledge. Look in well- peer-reviewed literature, as internet social comments are the most error-prone human communication that may have ever been devised.
It's not only the low-level dopaminergic stimulation from the blue light that zaps and dazes us, it's also the transcranial magnetic stimulation from the "low-level" radiofrequency that many parents want to pretend does not exist.
This is one of the best diagnoses of EdTech I've read this year, and it aligns with much of what I found while writing my own book on redesigning K-12 for the AI era.
The technology did not cause this. How it was used did. My research turned up the same pattern: schools bought devices without retraining teachers, without addressing the equity gaps in home broadband access, and without any real plan for which screens should and shouldn't replace. The result wasn't technology ruining kids. It was technology poured into a structure that had no theory of what it was for.
The fix can't be just removing devices, because the students Justine is teaching will graduate into a labor market that already runs on these tools. The employers doing the hiring aren't asking for kids who've never touched AI. They're asking for judgment: who can catch a tool being wrong, who can frame the right question, who can take responsibility for the output. That's a human skill, not a technical one, and it's exactly the kind of thinking a paper-first, handwriting-heavy, book-based early education builds. Protecting a six-year-old's attention span and preparing a seventeen-year-old to supervise AI responsibly at work are the same project on a different timeline.
That's the piece that gets lost when education and the labor market get treated as two separate conversations, one for parents and neuroscientists, one for economists and employers. A kid who never learned to sit with boredom, read a full book, or write by hand isn't just cognitively worse off- they are also less employable in the exact economy Winkleman describes, where the premium goes to people who can think, not people who can prompt. Schools that get the early years right and then deliberately teach supervision of AI and judgment as students get older are doing one job, not two.
This resonates deeply with what we've seen in Islamic schools trying to hold the line on the same pressures. The line about AI interrupting a fragile developmental process before it has a chance to unfold is exactly why we keep AI tools out of primary Islamic studies and Quran memorization entirely, a child's relationship with the text has to be built through repetition, recitation, and a human teacher's correction, not a chatbot's shortcut. Where we do use AI at all, it's on the administrative side of curriculum development at The Foundations Press (thefoundationspress.com), never in front of the student. Are you seeing any faith based school networks pushing back successfully, or is this fight mostly happening in secular public systems so far?
Recently it was suggested that even though a 14 yo is not on social media, the nature of the group text interactions causes a negative personality change. Take the phone away, and in about 4 days there is noticeable improvement. Fingers crossed.
I can’t be the only one who reads both this blog and some of the deeply-in-the-heart-of-AI ones like Astral Codex Ten.
What many of the true AI believers in SF/Silicon Valley think about education is actually scarier even than “screens help kids learn” or “jobs are all screen based anyway”. For many of them, they think that AI is a superhuman form of life that will soon (like, before 2030) take over most thinking and once AI itself designs better robots, most real world jobs too.
Half of them then think that humanity is cooked and the robots will kill us, and half think we’ll be kept as kind of like primates in a zoo. No need to learn to think or write or anything because what’ll be the point? The artificial superintelligence will do all of it better and faster anyway.
Even the tech-industry folks who don’t go this far still believe that little armies of AI agents will do all our thinking for us and take all actions, write your emails, etc. Also common is a belief that the thing to do is to become part of the capital-owning class as soon as possible since AI will supercharge wealth polarization even more than it already is.
All of which is to say — these people do not have our children’s best interests at heart. It all needs to be out of schools as soon as possible. More broadly: Our thinking is what makes us human. If we outsource that, what’s the point? (No wonder they don’t care about schools. They’ve given up on the future of the species, many of them.)
In developing knowledge of and experience with the use of AI, such as for research purposes, of how much of a machine it is, and a thinker it is not. In the end, a human should have to sign off on anything to propose it to be true.
I have surely seen the AI support a position, and then support the exact opposite position, and this can branch in multiple directions. I am reminded of what is needed to write a good research paper in 1950, or whenever, of that which has not changed. The best question I've ever asked an AI is, "What does the opposition say?" That is also the best question I have asked prior to the use of any AI. Tackling the topic from as many different angles as possible, and hunting down greater numbers of scenarios for context, with or without the machine, contributes the same quality to the result. Yes, the machine can wade through mountains of material in seconds, but only I can truly perform the high-level function, which will be to say that whatever is being said is ultimately true or false.
When you pull back the curtain of the screen, you are still left with a learning culture that has learned how not to learn. The learning culture is, and has always been, play deprived. That's why the temperament of young people is conditioned to gravitate towards the shiny things, the low-hanging fruit in the lives of everyone these days. The learning culture is a vacuum, a terrain of depleted soil where conversation could thrive if play propelled the classroom ecosystem the way it propels human intelligence. Play as a force of nature needs to be fully present in the hands and hearts of all the participants. An infrastructure composed of play -based communication is called for to make play central to the learning culture.
Hi Sophie. I worked in educational publishing for several decades. I came to it from the world of journalism. It was an absolutely wonderful experience that met a tragic ending: our company, an old and venerated one, was slowly destroyed by private equity. But that's another story. What surprised me was how fad-prone the field of education is. In the beginning, I would go to teachers conventions every year and every year the leaders in the field would be pushing another yet another new approach to teaching, often something involving the latest technologies. I grew very cynical pretty quickly and stopped going to the conventions.
Vert true. This fad cycle in education can be traced back more than a century (and is documented well in chapter 3 of “Education and Technology” by Neil Selwyn) with the rise and fall of educational film, radio, and TV. There are striking similarities between these technologies (which we would consider low-tech now) and today’s technologies. The problem is that none of them address root problems in education and rarely focus on the appropriate context to actually deliver education.,
Corporations have influenced schools for a very long time. By the late 1800s and early 1900s, wealthy industrialists began investing heavily in education.
Not because they funded everyday public schools directly through taxes—that remained primarily the responsibility of local and state governments—but because they shaped what schools taught and how they were organized. Major philanthropists included: Andrew Carnegie, John D. Rockefeller and Henry Ford. Their foundations financed:
Teacher training
Universities
School libraries
Vocational education
Educational research
Curriculum development
The Factory Model of Schooling
If we're still pretending that these philanthropists were simply trying to give back, we're being incredibly naive. Each of these men built their fortunes in industry, grooming immigrant adults and their entire families, how to behave in factory environment. This is when schools became organized around:
Bells
Fixed schedules
Grade levels
Standardized curriculum
Age grouping
Efficiency
Standardized testing
Large classrooms
The industrial model of education was never designed to cultivate independent, critical thinkers. It emerged during a period of rapid industrialization, when the economy demanded large numbers of workers who could reliably follow procedures, arrive on time, respect hierarchy, and perform standardized tasks. Schools increasingly mirrored factories: bells signaled transitions, students were grouped by age, learning was standardized, and success was measured by compliance and the ability to produce the "right" answers.
Critical thinking, questioning authority, and challenging accepted ideas were not the primary objectives of this system. In many ways, they were inconvenient. Imagine a factory floor filled with workers who constantly questioned every instruction from an authoritarian supervisor. Efficiency—the defining goal of industrial production—would grind to a halt. The educational system evolved to support the needs of that economy, rewarding obedience, conformity, and consistency far more often than curiosity, creativity, or independent thought.
That doesn't mean every educator wanted to suppress independent thinking. Many exceptional teachers have always encouraged students to ask questions, explore ideas, and think for themselves. But the structure of the system itself was built around standardization rather than intellectual exploration and capacity. The result is a legacy that still influences education today, even as the skills most needed in the modern world—critical thinking, adaptability, collaboration, and innovation—are the very qualities that standardized systems have often struggled to develop.
Many historians describe this as the "factory model." It mirrored industrial workplaces because that was considered the most efficient organizational system available. One well-known example was the Gary Plan, developed by William Wirt in the early 1900s.
Students rotated through:
Academic classrooms
Workshops
Gymnasiums
Auditoriums
This allowed schools to educate more students using the same buildings and reflected efficiency principles popular in industry.
Frederick Taylor and Scientific Management. The work of Frederick Winslow Taylor greatly influenced schools. Taylor developed "scientific management," which emphasized:
Standardization
Measurement
Efficiency
Specialized roles
Productivity
School administrators adopted many of these management practices, although classroom teachers often adapted or resisted them.
Were Schools Designed to Create Factory Workers? This is one of the most debated claims and as someone who has spent a fair amount of time in schools, I struggle with the idea that this ideas is even debatable at this point.
https://www.nytimes.com/2017/05/13/technology/google-education-chromebooks-schools.html
Schools were not created solely as factories for producing obedient workers. Industrial efficiency was one influence among several.
Who Actually Paid for Public Schools? The vast majority of funding has historically come from:
Local property taxes
State governments
Federal government (a much larger role after the mid-20th century)--also heavily influenced by corporations and their lobbyists.
Corporate "philanthropy" influenced educational policy and innovation and it's doing this today, pervasively.
Many features of today's schools reflect their 19th- and early 20th-century origins:
Age-based grades
School bells
Standardized testing
Fixed schedules
Uniform curricula
Seat-time requirements
If we hope to transform public education, we'll need to persuade the organizations with the greatest influence and resources—including major corporations—that investing in schools designed to cultivate critical thinking, creativity, and human potential is also an investment in their own future.
Tomorrow's workforce will not be defined by how well people follow instructions. Increasingly, success will depend on the ability to think critically, solve novel problems, collaborate across disciplines, and innovate in ways machines cannot. These are the very capacities that flourish through the arts, music, design, theater, and other creative pursuits that have too often been treated as expendable rather than essential.
The same is true for physical well-being. An active body supports an active mind. Research continues to demonstrate the connection between movement, exercise, attention, emotional regulation, learning, and mental health. Recess should not be viewed as wasted instructional time, and physical education should no longer be treated as an afterthought. They are foundational to helping children develop into healthy, engaged, resilient learners.
The irony is that the very skills and capacities many employers now say they desperately need—creativity, adaptability, emotional intelligence, collaboration, resilience, and sound judgment—are often the first casualties when schools become narrowly focused on standardized testing and measurable outputs. If corporations truly want a future workforce capable of navigating complexity and driving innovation, they have a vested interest in supporting an educational system that develops whole human beings, not simply efficient workers.
This is exactly the history the first article in my Foundations series traces, obedience, bells, standardized schedules, built for a factory economy rather than for critical thinking: https://newsletter.awarelife.co.il/p/the-system-works-thats-the-problem
Your closing point is the part worth sitting with longest, though. Persuading corporations to fund a different kind of education is a real, necessary fight, slow and structural, and it may take generations to actually shift. But a person doesn't have to wait for that fight to be won before they stop running the obedience pattern the system installed in them. That work can start now, in one person, regardless of what the system around them does or doesn't become.
Thanks for sharing. I take a different approach to evolution and change that doesn't start with a fight. The future is generated by inspiring vision--rather than fighting with what is. In terms of school transformation, that work is well under way in cities all across the US and in other places. Moving it to another level by including the stakeholders, (corporations), that benefit from more humans who have the skills and desire to think critically, is where we are now.
I don't think there's actual disagreement here either. Inspiring vision, bringing corporations and other stakeholders into a shared future, works at the level of the system. What I'm pointing at works at the level of the person, right now, inside whatever system currently exists.
Those aren't competing paths to the same destination. They're two different things happening at once. The system can be slowly reshaped by vision and inclusion, and a person can stop running an old pattern in themselves without waiting for that reshaping to finish. One doesn't have to fight the other for either to be true.
Totally agree!
I appreciate what you're saying. From the top of a big hierarchy, uniformity serves.
In an example from the article (LA), it took a citizen group to influence the school board in the right direction. More primary schools can be influenced more locally by parents/citizens. This is doable in an age where the general population appears to have no effect at a federal level. That would take national citizen organization - I think more likely, state-wide organization is doable.
Yes, I agree. Only, most parents these days, with both working, at least one job, maybe more, don’t have the space to dive into the PTA, or whatever it’s called these days, to advocate for these sense making changes. I’ve spent a fair amount of time working in environments with undeserved children and heard the comment that “these parents just don’t care” far too many times to stomach.
Robert Putnam has well-documented the loss of community. Perhaps we can make a list of ways in which forming local groups can benefit our health.
In a quick search, I did find this and will share that I agree with this 100%—When I bought my first laptop (early adopter over 30 years ago), I envisioned a time when more people would be working from home than going to the office. Who knew Covid was coming and the surge of remote workers. As someone who came out of the womb organizing social communities, I pondered that this would then require a heightened sense of community because human beings need socialization and if people were not engaged at work, they would need to create other spaces. Then came co-working spaces and an abundance of coffee shops that somewhat fills this need, but I haven’t seen either, and I’ve belonged to two co-working spaces that I enjoyed, but they’re really missing the mark on intentional community as they center only on workspace and ways to grow a business…. and coffee shops, at least where I live, miss the mark altogether and as a business consultant, that’s a strange dynamic to witness. Every shop and co-working space would profit more by taking this on…. “Robert Putnam's theory centers on social capital—the idea that social networks, norms of reciprocity, and trust hold communities together and make democratic societies function. He famously argued that a decline in civic engagement and community involvement leads to poorer governance, increased polarization, and widespread social isolation.”
Thanks for your note. I’m not familiar with Putnam’s work, but I do know that a lot of scholars and others have written about the loss of community and I have tons of my own theories as well. I like the thought of there being local groups that are helpful. Say more.
The main article on my 'stack is in 3 parts, and part 1 is an introduction to Putnam. Since that writing, from the AI there is also the idea of a CAPED group: Community Anchor for Participatory Ecosystem Development:
Asset Mapping & Shared Resource Management
Incubator for New Mutual Aid & Civic Groups
Community Concerns-to-Opportunities Pipeline
Rotating Leadership & Skill-Building
(Along with the idea in my article)
That's general thinking for now. As I say, I think we are getting into a position to think of making a list of health-related reasons for local groups.
Sophie, this is exactly right, and there's a mechanism underneath it worth naming. Learning isn't information transfer, it's the slow cultivation of a rich, personally-threaded memory substrate: the "fibrous journey" you describe. Handwriting, paper, boredom, a teacher's eyes on you: these all force the child to do the threading themselves, latent to active, effortfully.
Unfortunately, I currently teach math online mostly for the past 6 years since covid. Autocomplete or a chatbot tutor doesn't help the child weave their own mind together, it substitutes an external one. The engagement mechanics you name (streaks, rewards) aren't a design flaw bolted onto good pedagogy, they're what happens when a system optimizes for attention capture instead of memory formation. Because I teach online, I rely mostly on grading notebooks, with lots of practice with pencil and paper. This is my workaround. And project-based learning.
Bring back the boredom. Bring back the books.
Threaded Mind: Consciousness as Ecological Fabric and Recursive Memory Dynamics. https://doi.org/10.6084/m9.figshare.30505811
AI fingerprints are all over your comment. "there's a mechanism underneath it worth naming. Learning isn't information transfer, it's the slow cultivation of a rich, personally-threaded memory substrate: the "fibrous journey" you describe." There is irony in replying to a article about tech taking over schools...with obvious AI output.
We must restore critical thinking, actual memorization, handwriting, and collaboration. This is what I do with my students. "Fibrous": as in build our brains again through play and work and interactions with others. I long deeply for a day when AI chatbots are gone. I use them to edit my work. I make a lot of mistakes. But I'll be so happy when they are gone. There's a reason I still do not own a cell phone. There's a reason I'm reading this substack.
And why did big tech participate in undermining the education system to abandon the common sense that children cannot change their sex? Might attempts to change kids' exterior appearance with drugs and surgeries be a really bad idea that spreads like a contagion harming families and society on a mass scale?
Sophie — thank you for this essay. The engagement-metric slot-machine critique and the "personalized learning as depersonalized learning" quote from John Warner are two of the sharpest lines in the field this year, and I'll be citing both of you repeatedly.
For readers who stayed with the piece to the end: the products Sophie is describing — Chromebook autocomplete assistants, chatbot tutors on canonical content, engagement dashboards — are what I'd call controller architectures. AI deployed above the student, prescribing from outside. Complex Adaptive Systems theory of AI tutoring predicts they will fail exactly the way Sophie describes them failing: they fall into what my paper under review at BJET calls surface-compliance and mutual-capitulation attractors. Sophie's diagnosis is right; CAS names the mechanism.
There is a different architectural category — AI as constituent agent participating in a course's local rules, on a high-friction doer-effect substrate. My FIN 345 finance students at Ensign College compute CAMEL ratios by hand from raw 10-K filings; the AI persona works alongside them, not above them. It sustains the crawling phase Sophie rightly wants to protect.
Sophie's critique is required reading. The architectural alternative she may not have encountered is real, and it works.
— Chris Wasden, EdD
Your research on these types of AI sounds very interesting! What AI tool do your students use that works alongside them? Have you seen any other examples of AI as a “constituent agent”?
Brian — good questions, and I owe you honest answers on both.
The tool is Crucibl, the methodology and multi-agent architecture I've been building since 2024. It's currently deployed in FIN 345 (Financial Institutions) and FIN 485 (finance capstone). Four AI personas across the scaffold-fading arc — Parker (Full), Avery (Shared), Rowan (Light), Riley (faculty-facing) — each holding rules at the turn level: anti-fold discipline, completion-claim verification, audit-trail visibility. The RGAS architecture is specified in three papers currently under peer review. Public methodology overview at crucibl.tech.
On other examples: honestly, the explicit constituent-agent framing is not widely established yet, but there are adjacent efforts worth knowing about. Bastani et al. 2024 documented in a Turkey GPT-4 RCT that unstructured GPT-4 access produced 17% worse independent performance while a scaffolded configuration matched or beat the control group — the difference was architecture, not capability. Kestin et al. 2025 achieved Bloom two-sigma outcomes in a Harvard physics RCT with a structured tutor that resisted surface-compliance by design. The broader intelligent-tutoring-systems tradition (Kulik & Fletcher's 2016 meta-analysis) captures the pre-LLM version of the same architectural idea.
What's different in the Crucibl case is that the four failure modes are named explicitly and the rule-level countermeasures are architected in from the start. Whether that's the right way to frame the space is what peer review will tell me.
Thanks for reading, and for the good question.
— Chris Wasden, EdD
What a smart woman! She also happens to be a prominent member of the British Royal Family, yet unlike another actress member of that family who sells jam with her title, she never mentions it. That's because she is not an insecure arrivist.
Thank you Sophie for your powerful voice in this fight. EdTech and AI are also destroying the relationships that kids need with their teachers. “School connectedness,” which is driven by kids’ bonds with their teachers, not only promotes learning success but it also prevents kids from becoming depressed and suicidal. The last thing this increasingly depressed and suicidal generation of kids need is a screen that gets in the way of a real-world connection with a teacher. The first thing my kids talked about coming home from school was a story their teacher had told them. Such connections are what inspired them to do the hard work of learning. Every child deserves the opportunity to connect with a caring teacher, not an unfeeling laptop.
I think this is written by a knowledgeable person. I want to throw this angle in there. We have a rash of cases where there are emergency situations, and someone should dial 9/11, and people bust out there phones and film it instead. I see it looks like they view themselves as videographers, doing something useful, but also with later fallout, including either feeling shame in the following days for not actively participating, or by becoming more morally hardened.
In this sense, there is the question of what do screens do to the way we behave? We propose "cognitive offloading" in the case of information tools, which means we're not to remember anything. Plato didn't like the idea of writing for this reason, in a strongly oral time. I am also reminded of the US military's "information cockpit", which uses 7 screens that give context to the information being displayed, and increases memory retention by 55%, they say. I am reminded of reading a book, and remembering that something was halfway down the page on the left-hand side, maybe halfway through the book. That gives it a kind of context.
But screens are so much more of a way to fundamentally change our relationship with our world.
I believe, without checking that Plato merely recorded that Socrates advised against teaching writing.
He himself was literate. Socrates left no actual writings.
As i noted this way back in 9th grade i may be in error,but my memory ( moving often, i retain no Platonic materials) is that your comment concerning Plato is in error.
I asked the Google AI, and it said what's below. Perhaps this is also interesting in context of this article.
Plato argued that writing is bad for society because it weakens human memory, creates the illusion of wisdom, and destroys the interactive nature of true philosophical debate. He detailed these concerns in his dialogue Phaedrus through the myth of the Egyptian god Theuth (the inventor of writing) and King Thamus. [1, 2, 3, 4, 5]
Plato’s main critiques of writing include:
• It weakens memory: Plato argued that relying on written texts makes people forgetful because they stop exercising their natural memory. He viewed writing as a crutch rather than a tool for true learning. [1, 2]
• It lacks interactive dialogue: In a conversation, you can ask questions, clarify misunderstandings, and debate ideas (a process known as the dialectic). Plato believed written words are static; they cannot defend themselves if misunderstood or adapt to the reader's needs. [1, 2, 3, 4, 5]
• It creates the illusion of wisdom: Written text allows information to spread easily, giving readers the illusion that they are knowledgeable without actually going through the mental effort of understanding the concepts. [1, 2]
• It falls into the wrong hands: Writing cannot distinguish between who should and should not receive the information, broadcasting complex philosophical ideas indiscriminately to people who may not be ready for them. [1, 2]
Instead of writing, Plato championed living, spoken dialogue as the only way to achieve true philosophical enlightenment and deep understanding. [1, 2]
Children are NOT small adults. They are in the process of becoming!!! Some say until 25, others until 32. You listen to them not for wisdom but for freshness and surprise. Skipping the process parts leads to challenge conditions for that person
See the literature on dendritic pruning as well as that on PfC connectivity and myelinization.
Yes, 25 or so is the supposed final connective maturation.
The role of astroglia and other glia is being refined in neuroscience, and th full story of differential maturation , while outlined, is an important area of research.
We mature in quite distinct stages. See that well-known literature to help yourself become less tentative in comments.
Peripubertal adolescence , when salient attention largely shifts to peers is a VALUABLE stage in which adult-imposed cultural norms is and SHOULD be HIGHLY CRITICIZED by adolescents.
You will find that rational self interest is highest in that 13/4- 25 stage. After that submissiveness to obviously aberrant social structures and cultures, overcome actual intelligence.
I truncate this comment with the proviso that nearly all adults are, sometimes HIGHLY, deficient in understand what should be basic knowledge. Look in well- peer-reviewed literature, as internet social comments are the most error-prone human communication that may have ever been devised.
It's not only the low-level dopaminergic stimulation from the blue light that zaps and dazes us, it's also the transcranial magnetic stimulation from the "low-level" radiofrequency that many parents want to pretend does not exist.
Always ask: who benefits ?? It’s certainly NOT children or teachers!
This is one of the best diagnoses of EdTech I've read this year, and it aligns with much of what I found while writing my own book on redesigning K-12 for the AI era.
The technology did not cause this. How it was used did. My research turned up the same pattern: schools bought devices without retraining teachers, without addressing the equity gaps in home broadband access, and without any real plan for which screens should and shouldn't replace. The result wasn't technology ruining kids. It was technology poured into a structure that had no theory of what it was for.
The fix can't be just removing devices, because the students Justine is teaching will graduate into a labor market that already runs on these tools. The employers doing the hiring aren't asking for kids who've never touched AI. They're asking for judgment: who can catch a tool being wrong, who can frame the right question, who can take responsibility for the output. That's a human skill, not a technical one, and it's exactly the kind of thinking a paper-first, handwriting-heavy, book-based early education builds. Protecting a six-year-old's attention span and preparing a seventeen-year-old to supervise AI responsibly at work are the same project on a different timeline.
That's the piece that gets lost when education and the labor market get treated as two separate conversations, one for parents and neuroscientists, one for economists and employers. A kid who never learned to sit with boredom, read a full book, or write by hand isn't just cognitively worse off- they are also less employable in the exact economy Winkleman describes, where the premium goes to people who can think, not people who can prompt. Schools that get the early years right and then deliberately teach supervision of AI and judgment as students get older are doing one job, not two.
This resonates deeply with what we've seen in Islamic schools trying to hold the line on the same pressures. The line about AI interrupting a fragile developmental process before it has a chance to unfold is exactly why we keep AI tools out of primary Islamic studies and Quran memorization entirely, a child's relationship with the text has to be built through repetition, recitation, and a human teacher's correction, not a chatbot's shortcut. Where we do use AI at all, it's on the administrative side of curriculum development at The Foundations Press (thefoundationspress.com), never in front of the student. Are you seeing any faith based school networks pushing back successfully, or is this fight mostly happening in secular public systems so far?
Recently it was suggested that even though a 14 yo is not on social media, the nature of the group text interactions causes a negative personality change. Take the phone away, and in about 4 days there is noticeable improvement. Fingers crossed.