The Problem with The Anxious Generation — and What the “Ban All Screens” Movement Gets Wrong About Education

There is a real crisis in children’s mental health. I believe this because I see it daily in the schools I work in. The data supports it. The children themselves are telling us. On this point, Jonathan Haidt and I agree completely.

Where I part ways — and where I think the current panic about technology in schools is leading us somewhere counterproductive — is on the question of cause. And because cause determines response, getting this wrong has real consequences for real kids.

Let me be direct: The Anxious Generation is a compelling, well-written, emotionally resonant book built on a scientific case that is significantly weaker than Haidt presents it. The conclusions it has inspired in education policy are, in many cases, the wrong conclusions drawn from the wrong diagnosis — and I think educators and parents deserve a more honest accounting of where the evidence actually stands.


What Haidt Gets Right

Before the criticism, the credit.

Haidt is correct that something has gone badly wrong with childhood and adolescent wellbeing. He’s correct that overprotective parenting and the decline of play-based, independent childhood are serious problems. His advocacy for letting children take risks, experience failure, and develop resilience outside adult supervision — what he calls “antifragile” development — aligns closely with Peter Gray’s research and with what I see as an instructional coach working with students every day.

He’s also correct that smartphones and social media are not neutral tools for developing adolescents. The attention-capture dynamics, the social comparison mechanisms, the algorithmic amplification of outrage and anxiety — these are real design features with real effects. None of that is made up.

The problem is what he does with these legitimate observations. He builds an enormous causal argument on a foundation that the researchers who actually study this area largely reject.


The Scientific Case Against the Thesis

Candice Odgers, a developmental psychologist at UC Irvine, put it plainly in a review published in Nature: “The book’s repeated suggestion that digital technologies are rewiring our children’s brains and causing an epidemic of mental illness is not supported by science.” She added that the “bold proposal that social media is to blame might distract us from effectively responding to the real causes of the current mental-health crisis in young people.”

This isn’t one dissenting voice. The critics are numerous and credentialed. Andrew Przybylski, a professor of human behavior and technology at Oxford, describes Haidt’s approach as “vote counting” — prioritizing quantity of studies over quality, accumulating a long list of weak evidence and presenting it as a compelling case. Christopher Ferguson, a psychology professor at Stetson University who has studied media effects for decades, has pointed out that older adults in the US have experienced worse mental health deterioration than teenagers — which raises an obvious question: why would social media, used most heavily by the young, be causing problems worst in those who use it least?

One critical review examined the actual statistical rigor of the key studies Haidt relies on and found them wanting: “The book is over 400 pages long and waxes lyrical about the spiritual degradation we sustain as a result of social media… I would not have the nerve to write a several hundred page book calling for significant government intervention while summoning only five pages of statistical evidence. To make matters worse, the evidence is weak. The data quality is poor, the studies are flawed, and researchers are divided.”

The studies themselves have serious methodological problems. Many don’t study actual depressed teenage girls or heavy social media users — they study mostly adults, mostly average users, without serious psychological issues. You cannot establish the effect of heavy social media use on teenage depression unless you actually study heavy social media users who are depressed. Most of the studies Haidt cites don’t come close to that standard.


The Pattern I Keep Seeing

I grew up in the 80s and 90s. My generation was going to be ruined by television and video games. We were rotting our brains, becoming socially isolated, losing the capacity for deep attention and real connection. Parents panicked. Legislators proposed restrictions. Books were written explaining the neurological catastrophe underway.

Before my generation, it was comic books. Before that, rock music. Before that — and this is the one I find most useful to remember — novels. In the 18th and 19th centuries, novels were genuinely considered a moral hazard for young people, particularly young women. The idea that you would sit alone for hours, absorbed in a fictional world, engaging your imagination in ways that couldn’t be supervised or directed — this was seen as dangerous. Corrupting. The kind of thing that led to hysteria and bad decisions.

Every generation has a technological panic. The technology changes. The structure of the panic doesn’t. And the panic is always most persuasive to the people who didn’t grow up with the thing being panicked about. Ferguson draws a direct comparison to Seduction of the Innocent, the 1954 bestseller by psychiatrist Fredric Wertham that declared comic books had created a wave of juvenile delinquency — a book that caused enormous policy consequences before the evidence caught up with the panic.

I’m not saying social media is fine. I’m saying we’ve been here before, and the track record of these panics — as predictors of actual causal harm — is not good. The TV and video game generation didn’t turn out markedly worse than the generations before it. The novel-reading generation produced the Enlightenment.

What changes in each iteration is which thing we’ve decided is uniquely, irreversibly corrupting the youth. What doesn’t change is the confidence with which we assert it, the weakness of the actual evidence, and the policy consequences that follow before the evidence is properly interrogated.


What’s Actually Happening in Schools Right Now

The policy landscape has shifted fast. As of early 2026, some state legislators and witnesses have suggested banning 1:1 device programs in schools entirely, with calls for younger students to return to analog learning with pencil and paper. The Distraction-Free Schools Policy Project developed model legislation that would prohibit all screen technology in grades K-5 and ban school technology using generative AI at every grade level.

Parents across the country are forming networks teaching one another how to opt their children out of school-issued Chromebooks and iPads. One parent in California described pulling her children off school-issued devices as an “analog education” — framing it as a victory.

I understand the impulse. I genuinely do. Screen time management is a real issue. Distraction in the classroom is real. The feeling that technology has gotten away from us and we need to reclaim something is legitimate.

But the leap from “smartphones in pockets during class are a distraction” to “all screens in learning environments are harmful and we should return to pencil and paper” is enormous — and it’s a leap that the evidence doesn’t support.

Easier classroom management is not the same as better learning. And limiting students to pen and paper does little to prepare them for a world in which thinking, writing, and collaboration increasingly happen through digital tools.

There’s also an equity issue that gets papered over in these conversations. The children of affluent parents who are choosing analog education for their kids will still encounter a fully digital professional world. They’ll learn to navigate it eventually — at home, through tutors, through the social capital their families provide. The students who most need schools to close the digital literacy gap are the ones who will lose the most if we strip that from their education.


The Right Diagnosis, the Wrong Villain

Here’s what I think is actually happening, and why Gray’s framework matters more than Haidt’s for understanding it.

The mental health crisis in children is real and has been building since roughly the 1950s — decades before smartphones, social media, or the internet. Gray’s longitudinal data makes this undeniable. The primary driver, in Gray’s reading, is the progressive elimination of children’s independent, unstructured time: the reduction of recess, the increase in adult supervision, the overscheduling of childhood, the cultural shift toward treating independent children as negligent parenting.

Smartphones accelerated some of these dynamics and added new ones. But they arrived into a childhood that was already significantly impoverished of independent developmental experience. Children who have no free time, no unstructured outdoor play, no practice at self-regulation and conflict resolution — those children are developmentally primed for anxiety. Of course they reach for the nearest source of stimulation, connection, and escape. Of course the smartphone fills the vacuum.

The phone is a symptom as much as a cause. Taking the phone without restoring what the phone replaced is treating the symptom.

This is why I find the pencil-and-paper movement in education so frustrating. It’s addressing the wrong variable. A student who sits at a desk for six hours a day, goes home to an overscheduled afternoon of structured activities, and has never had two consecutive hours of genuinely unstructured time is not going to develop resilience because their school gave them a pencil instead of a Chromebook. The problem runs deeper than the device.


What Schools Should Actually Do

This is where I land, after years in the classroom and coaching teachers, watching students, and reading the research:

Cellphones during instructional time are a legitimate problem. Personal smartphones in pockets during class are a distraction issue, not a technology issue. Addressing that specifically — with clear policies, consistently enforced — is reasonable and has some evidence behind it.

1:1 device programs deserve scrutiny, but not abolition. The question isn’t whether devices belong in schools. It’s whether the learning design built around devices is pedagogically sound. The problem was never laptops. The real issue is the learning model we built around laptops. Bad technology implementation is a professional development and curriculum problem, not a technology problem.

The equity argument matters. Any policy that removes digital tools from schools disproportionately disadvantages students whose families can’t provide those tools and experiences at home.

Unstructured time is the real deficit. If we genuinely want to address the root causes of the mental health crisis Gray’s research describes, we need to give children back their unstructured time — at school and at home. More recess. Fewer scheduled activities. More space for boredom, conflict, and self-direction. That’s the intervention the data supports.

Teaching students to use technology critically is education, not capitulation. We live in a world saturated with algorithms designed to capture attention. The answer is not to pretend that world doesn’t exist or to seal children off from it until they turn 16 and then release them into it untrained. The answer is to help students develop the critical capacities to navigate it. That’s what education is for.


A Final Thought on Haidt

I’m not saying don’t read The Anxious Generation. It’s a book worth engaging with, and the parts of it that align with Gray’s research on free play and independent childhood are genuinely valuable. Haidt is a smart person thinking hard about a real problem.

But read it skeptically. Read the critics. Notice how much of the emotional weight of the book rests on anecdote and moral argument rather than the statistical case. Notice that the researchers who spend their careers studying this specific question — screen time and adolescent mental health — largely disagree with his conclusions.

And notice, most importantly, what the book makes it easy to avoid thinking about: the choices adults make about how to structure children’s time, how to design schools, how to build neighborhoods, how to value childhood independence. Those are harder conversations because they implicate us directly. Blaming the phone is easier. It usually is.


Further Reading

Free to Learn by Peter Gray — Start here. Gray’s full argument, written for a general audience, is rooted in decades of evolutionary psychology research. More compelling, better supported, and more actionable than anything else on this list.

Growing Up in Public: Coming of Age in a Digital World by Devorah Heitner — Published in 2023, this is the most current and most practically useful book on kids and technology that I’ve found. Heitner, a former media studies professor with a PhD from Northwestern, explicitly rejects the fear-based framing that dominates this conversation. Her core argument: the answer is mentoring, not monitoring. She draws on hundreds of interviews with kids, parents, and educators rather than extrapolating from weak correlational studies. A direct and well-earned counterweight to Haidt.

The Anxious Generation by Jonathan Haidt — Read it. Engage with the parts that align with Gray’s research on play deprivation. Push back hard on the causal claims about smartphones. It’s worth reading because it’s driving policy — and understanding the argument you’re pushing back against requires having read it.

Reclaiming Conversation by Sherry Turkle — Turkle is an MIT sociologist who has spent decades doing actual long-form qualitative research with students and families about technology and attention. More careful than Haidt, more specific about the mechanisms, and more interested in nuance than in producing a villain. Published in 2016, but holds up.

How Children Learn by John Holt — First published in 1967. Holt sat in classrooms, observed children learning — or not learning — and drew conclusions that the education system has ignored ever since. Gray cites him approvingly. The arguments about how children develop intrinsic motivation, curiosity, and self-direction are as relevant now as they were sixty years ago, possibly more so.



The Eclectic Educator is a free resource for everyone passionate about education and creativity. If you enjoy the content and want to support the newsletter, consider becoming a paid subscriber. Your support helps keep the insights and inspiration coming!

Yes, We Need to Get Rid of AP Courses

classmates doing studies for exam together
Photo by Armin Rimoldi on Pexels.com

And the College Board’s recent score inflation just made the argument stronger.


There, I said it. Let me make the case.

I’ve worked in public education long enough to watch AP courses go from a program for genuinely advanced students to a college admissions arms race to, now, something so thoroughly gamed by the College Board itself that universities are quietly questioning whether AP scores mean anything at all. We have spent the better part of two decades pushing AP as an equity solution — offering the best, most rigorous content to every student, regardless of background. That framing is correct. The vehicle we chose to deliver it is wrong.

Let me explain why, and what we should do instead.


The Equity Problem Has Never Been Solved

The original argument for expanding AP access was simple and appealing: if we give more students — low-income students, students of color, first-generation college students — access to rigorous coursework, we close the opportunity gap. More challenge equals better preparation equals better outcomes.

The data has never supported this at scale. A 2023 New York Times investigation found that roughly 60 percent of AP exams taken by low-income students scored too low for college credit — a 1 or 2 out of 5 — and that this number has barely moved in twenty years. Two decades of expanded access. Same failure rate. That’s not a pipeline problem. That’s a systemic problem with the model.

The barriers are layered and often invisible. Nationally, about 30 percent of Black and Hispanic students enrolled in AP courses never take the corresponding exam at all, compared to roughly 15 percent of Asian students. The reasons aren’t mysterious: scheduling conflicts, unofficial prerequisites, being steered toward “more appropriate” classes by counselors who read demographics rather than ability. Getting into the course doesn’t mean the course is actually accessible — or that success in it is equitably distributed.

This is the AP equity promise: a credential that most of the students it’s supposed to serve can’t access in any meaningful way.


The College Board’s Response: Change the Score, Not the System

Here’s where the story gets genuinely infuriating. After that NYT investigation put the failure rates for low-income and minority students into the national conversation, the College Board didn’t redesign courses, improve teacher training, or address structural barriers to preparation. They changed the scoring.

In 2022, the College Board quietly introduced what it calls “Evidence-Based Standard Setting” — a new methodology for scoring its most popular AP exams. The results were extraordinary, in the worst possible way.

AP U.S. History: students earning 4s and 5s jumped from 25 percent in 2023 to 46 percent in 2024. AP U.S. Government and Politics: top scores leapt from 24 percent to 49 percent in a single year. AP English Literature’s pass rate went from 44 percent in 2021 to 78 percent in 2022, the first year EBSS was applied.

Were students suddenly twice as prepared? Were teachers twice as effective? Did something happen in American high schools that would justify this kind of jump in a single year — while NAEP scores in 8th grade math and reading continued to decline and PISA scores showed stagnation or decline for American 15-year-olds?

No. The College Board changed the scoring system under pressure, and more students passed because passing got easier.

The financial context matters here. In 2024, over 86 percent of College Board revenue came from fees — nearly half of that from the basic AP exam fee alone. More than 1.3 million students paid $99 per exam for over 4.8 million AP exams in 2025. Total revenues exceeded $1.17 billion, and the organization held reserves of over $2 billion. The CEO received $2.3 million in total compensation in 2024 — comparable to the president of Stanford, whose institution operates on a budget roughly ten times larger.

The College Board has a direct financial incentive to keep AP attractive to students. If competitors like dual enrollment are growing, AP scores need to look competitive. The solution they chose wasn’t to improve the product. It was to make the grades better. Some elite universities are now quietly developing their own assessments to supplement AP data, having lost confidence in what AP scores actually signal.


What AP Courses Actually Do — and Don’t Do

Here’s the core problem, and it isn’t really about the College Board’s financial incentives, though those matter. It’s about what AP courses were designed to accomplish and what we’ve asked them to do instead.

AP courses were designed as an exam-prep system. The course exists to prepare students for the AP test. The test exists so students can demonstrate college-level knowledge and potentially earn college credit. That’s the whole loop. There’s nothing in that loop about authentic inquiry, personalized learning, or developing the kind of curiosity and self-direction that actually prepares people for college and life.

I’ve seen good teachers do extraordinary things inside AP courses. The structure doesn’t prevent great teaching — it just doesn’t require it, reward it, or build toward it. What it requires is covering the material on the exam. And teachers in underfunded schools, with overcrowded classrooms, serving students who haven’t had the preparation advantages their suburban peers have had, are left trying to jam college-level content into students who are already behind — while the clock ticks toward the May exam.

This is what we’ve decided counts as equity.

No one takes an AP course because it sounds exciting. Students take it because they need the credential, the weighted GPA boost, or the college credit — in roughly that order of priority. The course has become a box to check in a game nobody designed for the students who need the most from their education.


The Alternative That’s Already Working

Here’s what the advocates of the current system don’t want to talk about: dual enrollment is quietly eating AP’s lunch, and for good reason.

Dual enrollment allows high school students to take actual college courses — usually through community colleges or state universities — and earn real, transferable college credits before they graduate. Not maybe-credits that depend on a May exam score. Actual college credits that appear on an actual college transcript.

The numbers tell the story. In the 2024-25 school year, an estimated 2.8 million high school students were enrolled in dual enrollment courses — up from 2.5 million just two years earlier. Ninety percent of U.S. high schools now offer dual enrollment as of 2026. Studies consistently show that dual enrollment students are more likely to complete a bachelor’s degree, and the effect is particularly pronounced for first-generation college students.

Dual enrollment has real limitations. Quality varies by institution. Credit transfer isn’t guaranteed everywhere, particularly at highly selective universities. Some rural districts struggle with access to college partners. These are real problems worth solving.

But the structural difference matters enormously: in dual enrollment, the credit is earned by doing the work, not by performing on a single high-stakes exam in May. For students who’ve struggled all year and finally understood the material in April, AP rewards the exam. Dual enrollment rewards the semester.


What I Actually Want

I’m not just interested in replacing one credential with another. The deeper argument isn’t that dual enrollment is perfect — it’s that the entire framing of AP as an equity solution has distracted us from the real work.

The real work is redesigning Tier 1 instruction in every classroom for every student.

Not advanced placement for some. Not rigor for those who can access it through the right course label. Authentic, engaging, challenging learning environments for all students — where the goal isn’t coverage for an exam, but genuine intellectual development. Where teachers are supported and trained to create learning experiences that develop curiosity, critical thinking, and the capacity to learn independently. Where students who need more support get more support rather than being filtered into different tracks based on teacher recommendations and parental advocacy.

AP courses didn’t create tracking. But they reinforce it, give it a credential, and let us feel like we’ve addressed equity when the data says we haven’t.

As an instructional coach, I’ve watched schools celebrate expanding AP enrollment while the students enrolled in those courses received content coverage without the preparation, context, or support that would make it meaningful. The number of AP course offerings became a proxy for school quality. The number of students enrolled became a proxy for equity. The pass rates told a different story that nobody wanted to hear.

The College Board’s recent decision to fix that story by softening the scoring didn’t solve the problem. It made it harder to see.


The Hard Conversation

I know this argument is unpopular in certain circles. Parents who have watched their children use AP courses to build transcripts and earn college credit have real, concrete reasons to value the system. Teachers who’ve designed genuinely excellent AP courses have real, legitimate grievances with the suggestion that the whole structure should go.

I’m not saying those courses aren’t valuable. I’m saying the architecture around them — the College Board’s monopoly, the single high-stakes exam as the sole measure of learning, the financial incentives that led to score inflation, the equity promise that was never delivered — is worth being honest about.

We can do better. We should demand better. And the first step is being willing to say that a system that has failed its stated purpose for twenty years doesn’t deserve another twenty years of the benefit of the doubt.


Related on this site: The problem with The Anxious Generation and the “ban all screens” movement — a related argument about how education policy gets driven by compelling narratives rather than honest data.



The Eclectic Educator is a free resource for everyone passionate about education and creativity. If you enjoy the content and want to support the newsletter, consider becoming a paid subscriber. Your support helps keep the insights and inspiration coming!

What If Every Teacher Could Build an AI Tutor? David Wiley’s Generative Textbooks Idea Is Worth Your Attention

generative textbooks

There’s a particular kind of idea that shows up in education technology every few years — one that sounds almost too obvious once you hear it, but that nobody had quite put together that way before. David Wiley‘s work on generative textbooks is one such idea.

I’ve been following Wiley for a long time. If you’ve ever used an open textbook in a course or benefited from freely available educational materials online, there’s a good chance his fingerprints are on the infrastructure that made that possible. He’s one of the founders of the open educational resources movement — the effort to create, share, and freely adapt teaching and learning materials under open licenses. It’s unglamorous, important work that has saved students billions of dollars in textbook costs and given teachers genuine tools they can actually modify.

So when Wiley started applying that same philosophy to AI, I paid attention.


The Problem He’s Solving

The standard AI-in-education conversation goes like this: here are some tools (ChatGPT, Gemini, Claude, take your pick), and here are some ways teachers can use them. The tools belong to the companies. The teachers are users. If the company changes pricing, changes policy, or shuts down, the teacher starts over.

Wiley’s question is different: what if the instructional logic — the pedagogical intelligence built into an AI learning experience — belonged to the teacher? What if any educator could author an AI-powered learning tool without writing code, without a budget, and without surrendering control to a platform?

That’s what generative textbooks are attempting to answer.


How It Actually Works

The architecture is simpler than it sounds. A generative textbook isn’t a document — it’s a structured collection of inputs that, when assembled, tell an AI model exactly how to behave as a learning tool for a specific subject.

Here’s what an author creates:

  • A book-level prompt stub — the template that sets the AI’s voice, tone, format, and overall behavior. Think of this as the personality and ground rules of the learning experience.
  • Learning objectives — one per chapter or topic, short statements about what a learner should understand or be able to do.
  • Topic summaries — accurate, context-rich summaries written for the AI, not for students. These are what the model uses to stay grounded in accurate content rather than hallucinating.
  • Activity templates — the types of interactions available: flashcards, explanations, quiz questions, Socratic dialogue, whatever the author builds in.

When a student picks a topic and an activity type, the system assembles the relevant pieces into a single prompt and sends it to the language model, which generates a fresh, tailored learning experience — not retrieved from a database, but generated in the moment based on the author’s pedagogical structure.

As Wiley puts it: in this model, prompt engineering is instructional design. The authoring isn’t code — it’s curriculum work. That’s a meaningful distinction for teachers.


The Clever Pivot on Cost

The original prototype sent prompts through an API to open-weight language models hosted on Groq. Clean, seamless, technically elegant. Also not free — API calls cost money at scale, and Wiley found that most educators he consulted weren’t particularly concerned with whether the underlying model was “open” in the ideological sense. They were concerned with whether it was free for students.

So he made a pragmatic call: rather than routing prompts through a back-end service, the tool now assembles the prompt and copies it to the student’s clipboard. The student pastes it into whatever AI interface they already have access to — ChatGPT’s free tier, Gemini, a school-licensed model, whatever.

This is inelegant in the user-experience sense. There’s a copy-paste step that breaks the flow. Analytics become difficult. Student privacy depends on whatever tool they choose to use. Wiley is honest about all of this — he describes the project explicitly as a tech demonstration, not a finished product.

But there’s something worth noticing in the pragmatism. The decision prioritizes actual access over technical elegance. For students in districts that can’t afford platform licenses and teachers who don’t control their school’s technology budget, a tool that works with the free tier of a consumer AI product is more useful than a seamless experience behind a paywall.


Where Wiley Has Taken This Since

The generative textbook prototype was a starting point, and Wiley has kept building. His more recent thinking has evolved toward what he calls OELMs — Open Educational Language Models — a framework that combines open-licensed content with AI in a more sophisticated way.

The key addition is retrieval-augmented generation (RAG): rather than just grounding the AI’s behavior in a few paragraph-length topic summaries, an OELM includes a curated collection of OER content that the model actively retrieves from when generating responses. This makes the outputs more accurate, more traceable to specific source materials, and more trustworthy for educational use — one of the genuine limitations of relying on a general-purpose language model that might confabulate confidently.

The broader argument Wiley is making — that generative AI is the logical successor to OER — is worth sitting with. His claim isn’t that AI replaces open textbooks, but that the principles that made OER valuable (open licensing, participatory creation, the ability to adapt and remix) need to be extended into the AI space. As the educational materials market shifts toward AI-powered products, the question of who owns the instructional logic matters enormously for equity and access.


What This Means for Teachers

I want to be careful not to oversell where this project currently is. The generative textbooks site is live and explorable, but this is genuinely early-stage work. The copy-paste workflow has real friction. The quality of the learning experience depends heavily on the quality of the inputs a teacher creates, which means the authoring itself requires genuine pedagogical thought — garbage in, garbage out applies acutely here.

But the underlying question Wiley is raising is one I think about a lot as an instructional coach: who gets to design the learning experience, and on whose terms?

The dominant model in AI-powered education right now is platform-centric. A company builds an AI tool, schools license it, teachers become users. This mirrors exactly what happened with traditional educational technology — districts buy the LMS, teachers work inside it, the pedagogical architecture belongs to the vendor. We know how that story tends to go: cost escalation, lock-in, tools that don’t quite fit what teachers actually need because they were designed generically.

Wiley’s generative textbooks project is asking whether there’s another path — one where educators are architects rather than users. Where the instructional intelligence lives in open, adaptable, teacher-created structures rather than in proprietary platforms. Where a teacher in a school with limited resources can build a learning tool that’s as good as anything a well-funded district is paying for.

That’s not a modest ambition. And it’s not finished yet. But it’s the kind of work that tends to matter more than it seems to when it starts.


Go explore:


Related reading: my AI books post covers Ethan Mollick’s Co-Intelligence, which has useful framing for educators thinking about AI as a co-teacher rather than a replacement — a theme that runs directly through Wiley’s work.

AI as Co-Teacher or AI as Replacement?

bionic hand and human hand finger pointing
Photo by cottonbro studio on Pexels.com

“Empathy, evidently, existed only within the human community.”
— Philip K. Dick, Do Androids Dream of Electric Sheep?

There’s a moment in Philip K. Dick’s novel when the line between human and machine doesn’t shatter—it thins. The androids aren’t clumsy metallic caricatures. They’re articulate. Quick. Convincing. They can simulate emotional response so well that distinguishing them from humans requires careful testing. The danger isn’t brute force. It’s indistinguishability. It’s the subtle shift where simulation becomes “good enough,” and we stop asking what’s been replaced.

That’s what this moment in education feels like to me.

Not collapse. Not revolution. Just a quiet thinning of the line.

At Virginia Tech, a graduate course in Structural Equation Modeling nearly fell apart when the instructor unexpectedly dropped out. It was required. Students needed it to graduate. There wasn’t time to hire someone new. Instead of postponing the course, the department tried something that would have sounded like speculative fiction even five years ago. Half of the weekly learning objectives would be taught traditionally—through textbook and human instruction. The other half would be taught entirely through ChatGPT. Students received the same objectives either way. They completed the same assessments. And importantly, they submitted their AI chat logs along with their work so their reasoning could be examined. Every student passed.

You can read that as proof that AI can replace textbooks, maybe even instructors. Dr. Ivan Hernandez himself noted that AI can already function as a replacement for traditional textbooks and, to a certain extent, for instructors. That’s the easy interpretation, and it’s the one that will generate headlines.

But that’s not what interests me most.

What interests me is that Hernandez never surrendered the architecture.

He didn’t dissolve the classroom into a chatbot. He designed an experiment. He kept the objectives. He kept the assessments. He required documentation. He reviewed the logs. AI was allowed inside the system, but it did not define the system. The machine participated, but it did not govern.

That distinction feels subtle. It isn’t.

Because at the same time, another model of schooling is gaining attention. A 404 Media report on Alpha School states that students reportedly complete core academic work in roughly 2 hours per day. AI systems deliver most of the instruction. Adults function more as guides and coaches around the edges. The pitch is efficiency, personalization, and mastery at speed.

Now we’re standing inside the tension Dick was writing about decades ago.

If a system can simulate understanding, simulate responsiveness, and simulate personalized feedback, at what point do we stop asking whether it is human-centered?


When I talk about vibrant learning, I’m not talking about colorful classrooms or surface-level engagement. I’m talking about environments where students are actively constructing meaning, forming identity, navigating networks of knowledge, and experiencing the kind of belonging that makes intellectual risk possible. Vibrant learning is relational. It’s cognitively demanding. It depends on friction. It requires the presence of other minds.

And it is, almost by definition, inefficient.

The Science of Learning and Development has made something abundantly clear: learning isn’t merely cognitive processing. It is relational and contextual. Emotion and identity are braided into cognition. Belonging isn’t a nice add-on; it’s neurological infrastructure. When students feel safe enough to wrestle with ideas, they engage in deeper processing. When they feel unseen or disconnected, their cognitive system shifts toward protection rather than exploration.

Now imagine reorganizing schooling around algorithmic instruction as the primary academic engine.

Can AI explain structural equation modeling? Absolutely. The Virginia Tech experiment clearly demonstrates that. But explanation isn’t the same thing as formation. Learning is not just absorbing information; it’s situating yourself within a community of inquiry. It’s deciding what counts as credible. It’s learning how to disagree well. It’s building intellectual humility alongside intellectual confidence.

Connectivism adds another layer. Knowledge doesn’t reside in a single authority. It lives in networks—human, digital, and cultural. Learning is the ability to form and traverse those networks. AI belongs in that web. It can extend it. It can accelerate feedback loops. It can surface patterns that would take humans far longer to see.

But networks remain generative only when no single node dominates the topology.

When most academic interaction flows through a single algorithmic system, the structure centralizes. It becomes efficient. Predictable. Optimized. And optimization is not neutral. It always reflects a priority.

In Hernandez’s classroom, AI is one node among many. Students engage with it, but their interactions are documented and subject to human evaluation. The professor remains the architect. The AI is instrumentation. That’s augmentation.

In the Alpha-style model, as it’s been described, AI becomes the instructional spine. Humans support it. That’s substitution.

The difference between augmentation and substitution isn’t technological. It’s architectural.

And architecture shapes identity.


I understand why the efficiency model is appealing. Public education is strained. Teachers are exhausted. Districts are underfunded. Families are frustrated. If someone promises individualized instruction in two focused hours a day, it feels like relief. It feels like progress. It feels like the system finally catching up to the technology that already saturates students’ lives.

But we have to ask what we’re optimizing for.

If the goal is procedural mastery at scale, AI-centered instruction makes sense. You can compress problem sets. You can adapt pacing. You can automate feedback. You can produce measurable gains efficiently.

But public education, at its best, was never solely about workforce preparation. It was about citizenry. It was about forming people who can navigate complexity, ambiguity, disagreement, and shared life. That kind of formation doesn’t thrive in compressed, frictionless environments. It depends on relational tension. It depends on encountering other minds. It depends on spaces where empathy is not simulated but practiced.

Dick’s line lingers because it names something we’re tempted to overlook: empathy exists within the human community. Machines can model tone. They can generate encouragement. They can approximate responsiveness. But vibrant learning depends on something more than approximation. It depends on shared vulnerability, on the subtle cues of presence, on the unpredictable back-and-forth that shapes identity as much as it shapes understanding.

The Virginia Tech experiment shows that AI can assist with cognition. It does not prove that AI can replace the relational architecture in which cognition becomes character.

That’s the line.

It’s thin. And it’s easy to cross without noticing.

If pedagogy remains accountable to human judgment, AI can deepen vibrant learning. It can expand networks, accelerate iteration, and free educators to focus on the uniquely human dimensions of teaching. It can serve as a co-teacher inside a human-designed ecosystem.

But if pedagogy becomes accountable to platform architecture—if efficiency and throughput quietly become the organizing principles—then vibrant learning will slowly give way to optimized progression. The system may still function. Students may still perform. But something harder to measure will thin.

An educated workforce can be trained through efficient systems.

An educated citizenry must be formed within human communities.

The question before us isn’t whether AI works. It clearly does.

The question is who remains responsible for the architecture.

If we keep that responsibility—if we treat AI as instrumentation rather than architecture—then this moment could expand what’s possible in ways that genuinely support vibrant learning. If we don’t, if we reorganize schooling around efficiency engines and call it innovation, we may find that we’ve streamlined education while quietly narrowing what it means to be educated.

The machine can assist.

But empathy, formation, and responsibility still belong within the human community.

And whether that remains true in our schools will depend on the choices we make now—quietly, structurally, and often in the name of progress.



The Eclectic Educator is a free resource for everyone passionate about education and creativity. If you enjoy the content and want to support the newsletter, consider becoming a paid subscriber. Your support helps keep the insights and inspiration coming!

AI Schools and the Illusion of Efficiency

close up photo of an abstract art
Photo by Marek Piwnicki on Pexels.com

A recent investigation into Alpha School, a high-tuition “AI-powered” private school, revealed faulty AI-generated lessons, hallucinated questions, scraped curriculum materials, and heavy student surveillance. Former employees described students as “guinea pigs.”

That’s the headline.

But the real issue isn’t whether one school deployed AI sloppily.

The real issue is whether we are confusing technological acceleration with educational progress.

The Seduction of the Two-Hour School Day

Alpha’s pitch is simple and powerful: compress academic learning into two hyper-efficient hours using AI tutors, then free the rest of the day for creativity and passion projects.

If you believe traditional schooling wastes time, that promise is intoxicating.

But here’s the problem:

Efficiency is not the same thing as development.

From a Science of Learning and Development (SoLD) perspective, learning is not merely the transmission of content. It is a process that integrates cognition, emotion, identity, and social context. Durable learning requires safety, belonging, agency, and meaning-making.

You cannot compress belonging into a two-hour block.

You cannot automate identity formation.

And you cannot hallucinate your way to deep understanding.

Connectivism Is Not Automation

Some defenders of AI-heavy schooling argue that we are simply witnessing the next phase of networked learning. Knowledge is distributed. AI becomes a node in the network. Personalized pathways replace one-size-fits-all instruction.

That language sounds connectivist.

But Connectivism is not about replacing human nodes with machine ones.

It concerns the expansion of networks of meaning.

In a connectivist system:

  • Learning happens across relationships.
  • Knowledge flows through dynamic connections.
  • Judgment matters more than memorization.
  • Pattern recognition and critical filtering are essential skills.

AI can participate in that network.

But when AI becomes the primary instructional authority — generating content, generating assessments, evaluating its own outputs — the network collapses into a closed loop.

AI checking AI is not distributed intelligence.

It is recursive automation.

Connectivism requires diversity of nodes.

Not monoculture.

Surveillance Is Not Personalization

The investigation also described extensive monitoring: screen recording, webcam footage, mouse tracking, and behavioral nudges.

This is framed as personalization.

It is not.

It is optimization.

SoLD research clarifies that psychological safety and autonomy are foundational to learning. When students feel constantly watched, agency erodes. Compliance increases. Anxiety increases.

You can nudge behavior with surveillance.

You cannot cultivate intrinsic motivation that way.

If our model of learning begins to resemble corporate productivity software, we should pause.

Education is not a workflow dashboard.

The Hidden Variable: Selection Bias

To be fair, Alpha School reportedly produces strong test scores.

However, high-tuition schools serve families with financial, cultural, and educational capital. Research consistently shows that standardized test performance correlates strongly with income.

If affluent students succeed in an AI-heavy environment, that does not prove that the AI caused the success.

It may simply mean the students would succeed almost anywhere. I often say those students would succeed with a ham sandwich for a teacher.

The question is not whether AI can serve already advantaged learners.

The question is whether AI, deployed without deep pedagogical grounding, strengthens or weakens human development.

The Real Design Question

The danger is not AI itself.

The danger is designing educational systems around what AI does well.

AI does well at:

  • Drafting content
  • Generating practice questions
  • Scaling feedback
  • Recognizing surface patterns

AI does not do well at:

  • Reading emotional context
  • Building trust
  • Modeling intellectual humility
  • Navigating moral ambiguity
  • Forming identity

SoLD reminds us that learning is relational and developmental.

Connectivism reminds us that learning is networked and distributed.

If we optimize for what AI does well and marginalize what humans do uniquely well, we create a system that is efficient — but thin.

Fast — but shallow.

Impressive — but fragile.

What This Means for Public Education

This story is not merely about a private school engaging in aggressive experimentation.

It is a preview.

Every district will face pressure to:

  • Automate instruction
  • Replace textbooks with AI tutors
  • Compress seat time
  • Increase data capture

The answer cannot be a blanket rejection.

Nor can it be an uncritical adoption.

The answer is design discipline.

We should use AI to:

  • Reduce administrative drag
  • Prototype lessons
  • Support differentiated feedback
  • Expand access to expertise

But we should anchor every AI decision in two non-negotiables:

  1. Does this strengthen human relationships?
  2. Does this expand student agency and meaning-making?

If the answer is no, we are not innovating.

We are optimizing the wrong variable.

The Choice in Front of Us

We stand at a fork.

We can design AI systems around human development.

Or we can redesign human development around AI systems.

One path amplifies Connectivism, relational trust, and whole-child growth.

The other path creates compliant, monitored, hyper-efficient learners who score well but lack deep agency.

Technology will not make that choice for us.

We will.



The Eclectic Educator is a free resource for everyone passionate about education and creativity. If you enjoy the content and want to support the newsletter, consider becoming a paid subscriber. Your support helps keep the insights and inspiration coming!

Belonging Is a Design Choice

Belonging is one of the most talked-about—and most misunderstood—ideas in education.

We often treat it like a feeling that students either bring with them or don’t. If students feel disconnected, we respond with posters, slogans, or one-off activities meant to “build relationships.”

Those things aren’t bad. But they’re not enough.

Here’s the shift that matters:

Belonging isn’t a feeling problem. It’s a design problem.


What the Science of Learning and Development Tells Us

Research on the Science of Learning and Development (SoLD) shows that belonging is deeply linked to learning. Students are more likely to engage, persist, and take academic risks when they feel safe, seen, and valued.

But belonging doesn’t magically appear.

It’s shaped by instructional choices:

  • Who gets to speak—and how often
  • Who gets choice and agency
  • Whose knowledge and experiences are treated as valuable
  • How mistakes are responded to
  • Whether feedback invites growth or signals judgment

In other words, belonging lives inside the work itself.


Why Posters Aren’t Enough

A classroom can say “You belong here” on the wall and still send the opposite message through its design.

If tasks are rigid, voices are limited, and thinking is narrowly defined, students quickly learn where they stand.

Belonging isn’t something we add after instruction.

We build it into it.


Designing for Belonging

Designing for belonging doesn’t mean lowering expectations. It means creating structures that invite students to participate meaningfully.

That can look like:

  • Tasks with multiple entry points
  • Opportunities for students to connect learning to their experiences
  • Structured collaboration where every voice has a role
  • Feedback that focuses on growth instead of compliance

When belonging is intentional, students are more willing to engage deeply—and learning becomes more durable.


A Coaching Note from the Field

When teachers ask how to “build better relationships,” I often start with lesson design.

Relationships grow when students feel their thinking matters.

Belonging isn’t an add-on.

It’s an instructional choice.


If this way of thinking resonates, I write a short weekly newsletter for teachers and instructional leaders focused on authentic learning, instructional coaching, and designing schools that actually work.

You can subscribe here.

Engagement Is the Outcome, Not the Goal

For years, we’ve treated engagement like something teachers should be able to manufacture on demand.

If students aren’t engaged, the assumption is often that the lesson wasn’t exciting enough, interactive enough, or energetic enough. So we add activities. We add movement. We add tools. We add noise.

And then we’re surprised when it still doesn’t work.

Here’s the hard truth I’ve learned as an instructional coach:

Engagement isn’t something you plan for. It’s something you earn.


Why Planning for Engagement Often Backfires

When engagement becomes the primary goal of lesson planning, we usually end up designing around surface-level behaviors:

  • Are students busy?
  • Are they moving?
  • Are they talking?
  • Are they smiling?

But none of those things guarantees learning.

In fact, classrooms can look highly engaged while very little meaningful thinking is happening. Students comply. They complete. They perform school.

And teachers feel frustrated because they did everything “right.”


What the Research Actually Tells Us

Research connected to the Science of Learning and Development (SoLD) consistently points to the same conclusion:

Engagement follows meaning.

Students are more likely to engage when:

  • The task feels relevant to their lives or the world around them
  • They have some sense of ownership or choice
  • The thinking required actually matters

When those conditions are present, engagement emerges naturally. When they’re missing, no amount of energy can save the lesson.

This is why gimmicks don’t scale—and why they exhaust teachers.


Shifting the Planning Question

Instead of starting with:

“How do I make this engaging?”

Try starting with:

“Why would this matter to a student?”

That single question forces a different kind of design thinking:

  • What problem is being explored?
  • What decisions are students being asked to make?
  • Who or what is this work for?
  • Where does student thinking actually show up?

When lessons are built around those questions, engagement becomes a byproduct—not a burden.


What This Means for Teachers

This shift doesn’t require abandoning structure, rigor, or content. It requires recentering the work on meaningful thinking rather than performance.

It also reduces burnout.

When students carry more cognitive load, teachers don’t have to bring all the energy. The work itself does more of the heavy lifting.

That’s not about doing less—it’s about doing different.


A Coaching Note from the Field

When teachers tell me, “My students just aren’t engaged,” my response is rarely about strategies.

It’s usually about the task.

Fix the task, and engagement often surprises you.


If this way of thinking resonates, I write a short weekly newsletter for teachers and instructional leaders focused on authentic learning, instructional coaching, and designing school in ways that actually work.

No spam. No gimmicks. Just clear thinking from the field.

You can subscribe here.

MP Daily Telegraph: October 15, 2025

The Atlantic Telegraph 1866
The Atlantic Telegraph 1866 via Internet Archive
  • Illustrative Math’s CEO on What Went Wrong in NYC and Why Pre-K Math is Up Next – Illustrative Mathematics created a K-12 math curriculum used in many U.S. schools, but its rollout in New York City faced challenges due to implementation issues. The curriculum encourages students to think about problems before teachers explain solutions, blending direct teaching with student exploration. The organization is now focusing on early math by developing a pre-K curriculum to help students succeed from the start.
  • Mark Rober’s underwater search for a flooded Gold Rush mining town – (This is so FREAKING cool) Mark Rober used sonar and a small submarine to search for a flooded Gold Rush town under Folsom Lake in California. The town was covered by water after a dam was built in 1955. Despite challenges, the team found interesting shapes and objects on the lakebed.
  • D’Angelo: 14 Essential Songs – D’Angelo was a talented soul singer, songwriter, and producer known for his unique style and deep musicianship. He released three important albums blending soul, funk, jazz, and hip-hop, influencing the neo-soul movement. Despite personal struggles, his music remains powerful and full of emotion, exploring love, pain, and social issues.


The Eclectic Educator is a free resource for everyone passionate about education and creativity. If you enjoy the content and want to support the newsletter, consider becoming a paid subscriber. Your support helps keep the insights and inspiration coming!

A Quick Zine Resource Guide for Teachers

how to use zines with students

I’ve been on a zine kick for a while now, and recently had the chance to walk teachers through making their first zine.

We worked on creating their own zines, which was fun and made many of them uncomfortable, which is perfectly OK. I compiled some quick links and information, and we discussed potential ideas they might consider and run with when working with students.

Oh, and here’s the zine I made during one of the sessions. Feel free to use it to introduce the idea of zines to your peers and admin.

a zine about zines

Download the Zine About Zines

What Is a Zine?

  • A zine (short for “magazine” or “fanzine”) is a small-circulation, self-published work, often made by hand, that can take many forms—comics, essays, art, collages, instructions, etc.
  • Because zines are informal, tactile, and often DIY, they offer a low-stakes way for students to share voice, experiment with layout or narrative, and synthesize content in creative formats.
  • Zines are used in classrooms to teach skills such as media literacy, personal narrative, research synthesis, visual thinking, and more.

Folding a Zine — The One-Sheet Method

One of the simplest and most powerful forms is the one-sheet zine (fold, cut, fill).

Tools, Templates & Digital Zine Options

ResourceWhat It OffersLink / Notes
Zine-O-SphereSubstack exploring zines, art, culture, and DIY publishing.https://abigailschleifer.substack.com/s/zine-o-sphere 
“Using Zines in the Classroom and How to Make a Single Page Booklet Zine” (OER)Includes guidance + printable master flat for one-page zinesCUNY Academic Works
SCU Library’s Zine GuideWalkthroughs for physical & digital zines, plus design tips, templatesSCU Library Guides
The Arty Teacher: How to Make a ZineStep-by-step guide with photos, cutting/folding instructions, and classroom ideasThe Arty Teacher
“Teaching with Zines” (ZineLibraries.info)A compiled zine (yes, a zine) with resources, best practices, and reflections on using zines in educationzinelibraries.info
Barnard Zine Library – Lesson PlansSample lesson plans, ideas across content areas, ways to scaffold, suggestions for grading/feedbackzines.barnard.edu
TUIMP: The Universe In My Pocket“Using Zines in the Classroom and How to Make a Single-Page Booklet Zine” (OER)arXiv

Prophets of a Future Not Our Own

Photo by Zhimai Zhang on Unsplash
Photo by Zhimai Zhang on Unsplash

A friend made this prayer into a short video and, while the focus is on the work of Christians (real Christians, not the power-mad Christian Nationalists currently trying to ruin literally everything in the world), I can’t help but see our work as educators reflected here, as well.

This prayer was first presented by Cardinal Dearden in 1979 and quoted by Pope Francis in 2015. This reflection is an excerpt from a homily written for Cardinal Dearden by then-Fr. Ken Untener on the occasion of the Mass for Deceased Priests, October 25, 1979. Pope Francis quoted Cardinal Dearden in his remarks to the Roman Curia on December 21, 2015. Fr. Untener was named bishop of Saginaw, Michigan, in 1980.

It helps, now and then, to step back and take a long view.

The kingdom is not only beyond our efforts, it is even beyond our vision.

We accomplish in our lifetime only a tiny fraction of the magnificent
enterprise that is God’s work. Nothing we do is complete, which is a way of
saying that the Kingdom always lies beyond us.

No statement says all that could be said.

No prayer fully expresses our faith.

No confession brings perfection.

No pastoral visit brings wholeness.

No program accomplishes the Church’s mission.

No set of goals and objectives includes everything.

This is what we are about.

We plant the seeds that one day will grow.

We water seeds already planted, knowing that they hold future promise.

We lay foundations that will need further development.

We provide yeast that produces far beyond our capabilities.

We cannot do everything, and there is a sense of liberation in realizing that.

This enables us to do something, and to do it very well.

It may be incomplete, but it is a beginning, a step along the way, an
opportunity for the Lord’s grace to enter and do the rest.

We may never see the end results, but that is the difference between the master
builder and the worker.

We are workers, not master builders; ministers, not messiahs.

We are prophets of a future not our own.



The Eclectic Educator is a free resource for everyone passionate about education and creativity. If you enjoy the content and want to support the newsletter, consider becoming a paid subscriber. Your support helps keep the insights and inspiration coming!