A neuroscientist’s Senate testimony has been viewed over two million times. Parents are showing up to school board meetings with copies of his book. Sixteen states have introduced legislation to restrict devices in schools.
The four pillars are useful. But "purpose before platform" is asking a question the procurement process already answered. Districts don't sign contracts after teachers decide why the tool serves the learning. They sign contracts, then ask teachers to find a use. The structural failure is that the decision architecture doesn't require an educational rationale before it requires a signature.
The most important data point in this whole piece is near the end: disadvantaged students got the highest doses of edtech and the weakest outcomes. That's the system doing what it was designed to do: moving product through the schools with the least resistance. No four-pillar framework fixes that upstream.
Completely agree, Christopher. The decision architecture is backwards and one of the primary reasons so many schools are left with digital things and not purpose driven technology.
The same goes with adults, which is why I published [_Cognitive Productivity: Using Knowledge to Become Profoundly Effective_](https://leanpub.com/cognitiveproductivity/) and [_Cognitive Productivity with macOS: 7 Principles for Getting Smarter with Knowledge_](https://leanpub.com/cognitive-productivity-macos), to provide knowledge workers with the cognitive science and the skills they need to become more cognitively productive with IT.
I really appreciate this perspective. Our group just took a petition to our school board in Seattle asking for more transparency and intention behind EdTech decisions.
I did just hear a podcast interview with Horvath where he says he’s not anti edtech but wants to see data supporting its effectiveness above analog methods to justify its adoption. I think he actually poses questions very similar to yours though maybe a little more stringent. And I shared your post with a friend and said it would be super informative to see a collegial dialogue between the two of you about this where you differ though I think you’re saying very similar things.
And as a parent of kids who will be in school for the next fourteen years, I would prefer things be yanked back and readded thoughtfully with well supported teachers than continuing with minuscule tweaks while kids spend unknown amounts of time with gamified apps and YouTube videos. It feels so insidious and the arguments like “kids need to know how to use tech in this world” are woefully empty of real critical thinking and designed to shut down conversation as opposed to considering ideas like your pillars or Horvath’s questions.
Erica, thank you — and seriously, kudos for taking that petition to the board. That kind of parent-led pressure is exactly what shifts how districts make these decisions.
You're right that Horvath and I are circling the same questions. I'd genuinely welcome that dialogue. His instinct to demand evidence before adoption is the right one; I tend to land on "intentional integration with well-supported teachers" as the path forward, but the diagnosis is the same — too much tech has been rolled out on vibes rather than evidence.
And your last point lands hard. "Kids need to know how to use tech in this world" has become the conversation-ender, and it's intellectually lazy. Knowing how to use tech well is precisely what we're not teaching when we hand kids gamified apps and call it learning.
👏👏👏 No notes. As a complete tech geek English teacher, I actually don’t deploy the 1:1 chromebooks we have that much. Thinking before typing. Also the exams are still handwritten so there’s that…
I ALSO have made Gems, regularly set assignments in google classroom, bounce around collab docs, share NotebookLMs, use JSTOR AI, MassoLit AI, Brisk…
Nice essay. You write: “Intelligent tutoring systems — where technology adapts to what a student knows and doesn’t know — show strong positive effects (+0.52). Interventions designed for students with learning differences show even stronger ones (+0.61).”
I think there's something just "off the screen" not addressed here: the fact that in most cases, kids can be using a great learning product, but are a single clip or swipe away from something that's very off-task (at best) and harmful (at worst). That's another part of the device issue that is FAR too often unaddressed (and understandably so, because it's a wicked problem)
Thank you for this. I brought the statistical incoherence of the argument directly to Horvath, only for him to call me juvenile and tell me I act like a PhD student lol. It’s appalling that someone that’s irresponsibly crafted an argument that spans a whole book is moving the needle so much and fomenting moral panic. Any argument that swings the pendulum too far in one direction is most probably a straw man!
I think this is a fair pushback against the idea that all classroom technology is the same thing or that correlation alone proves devices caused declining outcomes.
But I also think “purposeful use” can become a bit too easy of an answer. In actual classrooms, a screen activity can look active and engaging while still producing less thinking, less conversation, less self-regulation, or less durable learning than the non-screen option.
For me, technology should not be the default. It should have a clear purpose and earn its place: does it give students access, feedback, creative possibilities, or learning we could not reasonably get otherwise? If not, I would rather protect time for books, handwriting, making, discussion, and sustained attention, especially with younger students.
So I am not anti-tech. I just think the burden should be on us to show why the device is the better tool for this learning, right now.
You describe Horvath's ideas as in opposition to your own, yet I think this is disingenuous. Having read "The Digital Delusion" and watched all of his Cognitive Science and Metacognition videos, asserting that his prescription is to "pull the plug" seriously misconstrues his arguments.
I'm the Senate testimony you link, he states in his conclusion "This is not a debate about rejecting technology. It is a question of aligning educational tools with how human learning actually works. "
Additionally, The four pillars you describe align very neatly with his suggestions at the conclusion of "The Digital Delusion." You emphasize "purpose before platform", and on page 186 he suggests teachers use the following questions to reflect on digital technology use in their classroom:
What tools are being used?
By whom?
For what purpose?
For how long?
Your pillar of "Active over Passive" matches his suggestion that students be restricted from access to autocomplete and autocorrect tools, so as to incentivize active recall (pg 162). These tools do not engage retrieval from long-term memory and therefore do not promote information retention (pg 187). As an aside, I feel it important to note that "active learning" is poorly defined in that physical activity and verbal activity are often what educational professionals look for when surveying a classroom for "engagement", whereas cognitive activity is not so easily observed In a classroom context, for obvious reasons. That is, "minds on" learning is what we should all be striving for as teachers, regardless of the medium through which the learning occurs. Paul Kirschner discusses the illusions of student engagement as an indicator of learning on Anna Stokke's podcast, Episode 56.
For the third pillar, You suggest putting the teacher first. On page 234 of his book Horvath has a subheading titled " recenter the teacher" , where he emphasizes the importance of schools celebrating and supporting their teachers.
The fourth pillar you suggest is in exact alignment with the entirety of his last 50 pages, which emphasizes open dialogue between students, parents, educators, and school leaders about a "balanced approach that reflects evidence-based practice, equity and the values of community." His book includes actual parent action toolkits, coalition action menus, and drafts of potential letters and initiatives for parent teacher coalitions.
I'm summary, his book is in near total alignment with your philosophy, and if you read it you will find a wealth of ideas similar and more fully fleshed out than yours.
As for AI tutors and that +0.52 effect that keeps being touted, I remain highly skeptical. On page 190 Horvath reminds us: "screens impose an extremely narrow context. When all learning happens on a device that device becomes part of every episodic memory-and eventually, every semantic one." Research is already coming out that reinforces all of my reservations about AI tutors : https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6868618
Thank you for this — genuinely. This is exactly the kind of dialogue I wish we had more of in this debate, and you've clearly done the reading. I appreciate the care you put into it.
You're right that there's real overlap between Horvath's concluding recommendations and the framework I've been building. Honestly, that's encouraging, not threatening. If someone coming from the cognitive science side and someone who's spent two decades running technology programs inside schools arrive at similar principles — purpose before platform, recentering the teacher, open dialogue with families — that's a signal we're circling something true. I don't need to own those ideas. Nobody does. They're older than both of us.
Where I push back isn't on the fine print — it's on the framing. When the nuance lives on page 186 and the title says "delusion," the title is what travels. The Maine piece is my case in point: he concedes twice that the charts prove nothing causal, then lets them do the persuading anyway. That rhetorical move feeds a policy wave that is far blunter than his own conclusions — school boards don't legislate from page 234, they legislate from headlines. My frustration is with what the argument does in the world, and with how rarely the practitioners who'd have to live with those policies are in the room.
On "minds on" learning — full agreement, and thanks for the Kirschner pointer. Visible busyness is not cognition, and edtech has been guilty of confusing the two for years. I've said as much about our side of the aisle.
And on AI tutors: I share more of your skepticism than you might expect. The same discipline I'm asking of Horvath — don't let one dramatic effect size carry more weight than its context can bear — applies equally to the +0.52 claims. I'll read the paper you linked.
Very thought-provoking. One thing that often gets left out of the conversation is some children with special needs for whom technology has absolutely opened doors to communicating and learning that were never there before. But, what is life-changing for one student could be counterproductive to another. We need an educational system that is not one size fits all.
Andrew, this piece gave me the most important context I've seen in this entire conversation around Horvath's testimony. The distinction you drew between the OECD data measuring leisure screen time versus purposeful, teacher-guided learning is a vital detail that keeps getting stripped out in every viral clip and school board panic and it completely changes the false conclusion people are drawing.
I just dropped a LinkedIn post today pushing back on Sweden's decision to retreat to paper, and your work was central to making the argument accurately. I quoted you directly and made sure Horvath's actual position wasn't misrepresented because the gap between what he said and what people think he said is doing real damage to this conversation.
If you get a chance to read it I'd genuinely value your thoughts especially on whether the AI-in-classrooms warning at the end resonates with what you're seeing in schools right now.
The four pillars are useful. But "purpose before platform" is asking a question the procurement process already answered. Districts don't sign contracts after teachers decide why the tool serves the learning. They sign contracts, then ask teachers to find a use. The structural failure is that the decision architecture doesn't require an educational rationale before it requires a signature.
The most important data point in this whole piece is near the end: disadvantaged students got the highest doses of edtech and the weakest outcomes. That's the system doing what it was designed to do: moving product through the schools with the least resistance. No four-pillar framework fixes that upstream.
Completely agree, Christopher. The decision architecture is backwards and one of the primary reasons so many schools are left with digital things and not purpose driven technology.
Just a minor correction - you say "A 2024 study published in ScienceDirect..."
ScienceDirect is a platform for accessing journals, not a journal. The study was published in Complementary Therapies in Medicine.
Very thoughtful article!
Well, you’re still smart. Good article.
The same goes with adults, which is why I published [_Cognitive Productivity: Using Knowledge to Become Profoundly Effective_](https://leanpub.com/cognitiveproductivity/) and [_Cognitive Productivity with macOS: 7 Principles for Getting Smarter with Knowledge_](https://leanpub.com/cognitive-productivity-macos), to provide knowledge workers with the cognitive science and the skills they need to become more cognitively productive with IT.
I really appreciate this perspective. Our group just took a petition to our school board in Seattle asking for more transparency and intention behind EdTech decisions.
I did just hear a podcast interview with Horvath where he says he’s not anti edtech but wants to see data supporting its effectiveness above analog methods to justify its adoption. I think he actually poses questions very similar to yours though maybe a little more stringent. And I shared your post with a friend and said it would be super informative to see a collegial dialogue between the two of you about this where you differ though I think you’re saying very similar things.
And as a parent of kids who will be in school for the next fourteen years, I would prefer things be yanked back and readded thoughtfully with well supported teachers than continuing with minuscule tweaks while kids spend unknown amounts of time with gamified apps and YouTube videos. It feels so insidious and the arguments like “kids need to know how to use tech in this world” are woefully empty of real critical thinking and designed to shut down conversation as opposed to considering ideas like your pillars or Horvath’s questions.
Erica, thank you — and seriously, kudos for taking that petition to the board. That kind of parent-led pressure is exactly what shifts how districts make these decisions.
You're right that Horvath and I are circling the same questions. I'd genuinely welcome that dialogue. His instinct to demand evidence before adoption is the right one; I tend to land on "intentional integration with well-supported teachers" as the path forward, but the diagnosis is the same — too much tech has been rolled out on vibes rather than evidence.
And your last point lands hard. "Kids need to know how to use tech in this world" has become the conversation-ender, and it's intellectually lazy. Knowing how to use tech well is precisely what we're not teaching when we hand kids gamified apps and call it learning.
👏👏👏 No notes. As a complete tech geek English teacher, I actually don’t deploy the 1:1 chromebooks we have that much. Thinking before typing. Also the exams are still handwritten so there’s that…
I ALSO have made Gems, regularly set assignments in google classroom, bounce around collab docs, share NotebookLMs, use JSTOR AI, MassoLit AI, Brisk…
It. Does. Not. Have. To. Be. A. Binary.
Nice essay. You write: “Intelligent tutoring systems — where technology adapts to what a student knows and doesn’t know — show strong positive effects (+0.52). Interventions designed for students with learning differences show even stronger ones (+0.61).”
Might I trouble you for your sources on this?
Well said--thanks for writing this up.
I think there's something just "off the screen" not addressed here: the fact that in most cases, kids can be using a great learning product, but are a single clip or swipe away from something that's very off-task (at best) and harmful (at worst). That's another part of the device issue that is FAR too often unaddressed (and understandably so, because it's a wicked problem)
Thank you for this. I brought the statistical incoherence of the argument directly to Horvath, only for him to call me juvenile and tell me I act like a PhD student lol. It’s appalling that someone that’s irresponsibly crafted an argument that spans a whole book is moving the needle so much and fomenting moral panic. Any argument that swings the pendulum too far in one direction is most probably a straw man!
I think this is a fair pushback against the idea that all classroom technology is the same thing or that correlation alone proves devices caused declining outcomes.
But I also think “purposeful use” can become a bit too easy of an answer. In actual classrooms, a screen activity can look active and engaging while still producing less thinking, less conversation, less self-regulation, or less durable learning than the non-screen option.
For me, technology should not be the default. It should have a clear purpose and earn its place: does it give students access, feedback, creative possibilities, or learning we could not reasonably get otherwise? If not, I would rather protect time for books, handwriting, making, discussion, and sustained attention, especially with younger students.
So I am not anti-tech. I just think the burden should be on us to show why the device is the better tool for this learning, right now.
You describe Horvath's ideas as in opposition to your own, yet I think this is disingenuous. Having read "The Digital Delusion" and watched all of his Cognitive Science and Metacognition videos, asserting that his prescription is to "pull the plug" seriously misconstrues his arguments.
I'm the Senate testimony you link, he states in his conclusion "This is not a debate about rejecting technology. It is a question of aligning educational tools with how human learning actually works. "
Additionally, The four pillars you describe align very neatly with his suggestions at the conclusion of "The Digital Delusion." You emphasize "purpose before platform", and on page 186 he suggests teachers use the following questions to reflect on digital technology use in their classroom:
What tools are being used?
By whom?
For what purpose?
For how long?
Your pillar of "Active over Passive" matches his suggestion that students be restricted from access to autocomplete and autocorrect tools, so as to incentivize active recall (pg 162). These tools do not engage retrieval from long-term memory and therefore do not promote information retention (pg 187). As an aside, I feel it important to note that "active learning" is poorly defined in that physical activity and verbal activity are often what educational professionals look for when surveying a classroom for "engagement", whereas cognitive activity is not so easily observed In a classroom context, for obvious reasons. That is, "minds on" learning is what we should all be striving for as teachers, regardless of the medium through which the learning occurs. Paul Kirschner discusses the illusions of student engagement as an indicator of learning on Anna Stokke's podcast, Episode 56.
For the third pillar, You suggest putting the teacher first. On page 234 of his book Horvath has a subheading titled " recenter the teacher" , where he emphasizes the importance of schools celebrating and supporting their teachers.
The fourth pillar you suggest is in exact alignment with the entirety of his last 50 pages, which emphasizes open dialogue between students, parents, educators, and school leaders about a "balanced approach that reflects evidence-based practice, equity and the values of community." His book includes actual parent action toolkits, coalition action menus, and drafts of potential letters and initiatives for parent teacher coalitions.
I'm summary, his book is in near total alignment with your philosophy, and if you read it you will find a wealth of ideas similar and more fully fleshed out than yours.
As for AI tutors and that +0.52 effect that keeps being touted, I remain highly skeptical. On page 190 Horvath reminds us: "screens impose an extremely narrow context. When all learning happens on a device that device becomes part of every episodic memory-and eventually, every semantic one." Research is already coming out that reinforces all of my reservations about AI tutors : https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6868618
Thank you for this — genuinely. This is exactly the kind of dialogue I wish we had more of in this debate, and you've clearly done the reading. I appreciate the care you put into it.
You're right that there's real overlap between Horvath's concluding recommendations and the framework I've been building. Honestly, that's encouraging, not threatening. If someone coming from the cognitive science side and someone who's spent two decades running technology programs inside schools arrive at similar principles — purpose before platform, recentering the teacher, open dialogue with families — that's a signal we're circling something true. I don't need to own those ideas. Nobody does. They're older than both of us.
Where I push back isn't on the fine print — it's on the framing. When the nuance lives on page 186 and the title says "delusion," the title is what travels. The Maine piece is my case in point: he concedes twice that the charts prove nothing causal, then lets them do the persuading anyway. That rhetorical move feeds a policy wave that is far blunter than his own conclusions — school boards don't legislate from page 234, they legislate from headlines. My frustration is with what the argument does in the world, and with how rarely the practitioners who'd have to live with those policies are in the room.
On "minds on" learning — full agreement, and thanks for the Kirschner pointer. Visible busyness is not cognition, and edtech has been guilty of confusing the two for years. I've said as much about our side of the aisle.
And on AI tutors: I share more of your skepticism than you might expect. The same discipline I'm asking of Horvath — don't let one dramatic effect size carry more weight than its context can bear — applies equally to the +0.52 claims. I'll read the paper you linked.
Thanks for engaging!
Thanks for making these important pts!
Thanks, Michael!
Very thought-provoking. One thing that often gets left out of the conversation is some children with special needs for whom technology has absolutely opened doors to communicating and learning that were never there before. But, what is life-changing for one student could be counterproductive to another. We need an educational system that is not one size fits all.
Andrew, this piece gave me the most important context I've seen in this entire conversation around Horvath's testimony. The distinction you drew between the OECD data measuring leisure screen time versus purposeful, teacher-guided learning is a vital detail that keeps getting stripped out in every viral clip and school board panic and it completely changes the false conclusion people are drawing.
I just dropped a LinkedIn post today pushing back on Sweden's decision to retreat to paper, and your work was central to making the argument accurately. I quoted you directly and made sure Horvath's actual position wasn't misrepresented because the gap between what he said and what people think he said is doing real damage to this conversation.
If you get a chance to read it I'd genuinely value your thoughts especially on whether the AI-in-classrooms warning at the end resonates with what you're seeing in schools right now.
https://www.linkedin.com/posts/giwahenry_a-neuroscientist-stood-in-front-of-the-us-share-7451917581763944448-lg2P?utm_source=share&utm_medium=member_desktop&rcm=ACoAAD2adoYBNaRrN3AwO0NwAzgGP9Ppv_0zAvs