This essay first appeared in my newsletter Sign up here if interested in Unf^cking Education.

---

Why AI in schools, as currently imagined, won’t fix what’s broken.

This month’s executive order to bring AI into K–12 education sounds exciting and bold. A national task force. A presidential challenge. Public-private partnerships. Federal dollars. Teacher training. Early exposure.

It checks all the boxes.

Except the ones that matter.

Because for most schools, this won’t be a revolution. 

• It’ll be another slide deck or memo. 
• Another PD (professional development) session. 
• Another initiative. 
• Another set of tech vendor demos. 
• Another thing piled onto overburdened teachers who haven’t been given the time, tools, or trust to make change stick.

AI isn’t going to save our schools. It’s going to expose how broken they already are.

And this isn’t a cynical take. 

It’s just pattern recognition.

The Cycle We Can’t Seem to Break

We’ve seen this before—technological solutions cycling through schools without changing fundamentals aka educational bread & circuses:

• iPads in every classroom (2010-2013): Districts spent billions, Apple touted “educational revolution,” teachers got one-day training, and most ended up as expensive PDF readers.
• SMART Boards (2000s): Schools replaced blackboards with $5,000 interactive displays that quickly became projector screens when the calibration drifted and IT couldn’t keep up.
• Hour of Code (2013-present): Annual PR event that convinced parents programming was happening while actual computer science education remained unchanged.
• STEM/STEAM initiatives (2010s): Rebadged existing science classes, added 3D printers few knew how to use, changed nothing about how science was taught.
• Personalized Learning Platforms (2015-present): Promised adaptive instruction, delivered glorified multiple-choice tests with progress bars.
• SEL Curriculum (2018-present): Added posters about emotions to classrooms where teachers still had no time for meaningful student connections.

Each rollout promised transformation but delivered tokenism. The tools changed, but the incentives and classroom practices didn’t.

And the results continue to get worse. 

We still have two-thirds of high school students disengaged. 

Reading and math scores are declining. Critical thinking is rare. Writing is worse. Teachers are burning out. And students are wondering, more than ever, when will I ever use this?

This is the context AI is walking into.

America vs China

This executive order is a reflexive response to China’s national AI education push and the ensuing hysterical headlines (see below). 

It is not a thoughtful approach based on classroom realities. It’s the equivalent of buying a fancy car because your neighbor bought one. It is about creating the appearance that America hasn’t fallen behind in the global AI race.

The biggest irony here is that students are already far ahead of policymakers—they’ve been using AI for months  to circumvent the pointless assignments we keep giving them. While we write memos about “AI literacy,” students are using ChatGPT to generate five-paragraph essays about books they haven’t read and solving math problems they don’t understand.

Making ineffective education more efficient isn’t progress. 

It’s like upgrading from a rowboat to a speedboat when you’re headed for Niagara Falls. You’ll reach the precipice faster, with fancier equipment, but you’re still going over the edge.

Speed is great, but what we really need is a compass.

Missing from the Executive Order: Reality

The steelman case for AI in schools is strong. 

Used well, AI could personalize instruction, help students build real-world tools, free teachers from soul-killing admin work, offer feedback at scale, and teach data literacy and automation skills.

But the order creates more burdens without addressing the fundamental constraints that kill innovation in schools:

• Time Poverty: Teachers already work 50+ hour weeks and are horribly paid. When will they learn, experiment with, and implement AI tools?
• Initiative Fatigue: Schools continuously cycle through programs without seeing any through. Most teachers have learned to wait out each new thing.
• Misaligned Incentives: Schools are judged by test scores and compliance metrics, not innovation or student agency.
• Tech Infrastructure Gap: Many schools still struggle with basic internet reliability and device access.
• Evaluation Theater: Success will be measured by “implementation statistics” (how many teachers attended trainings) rather than meaningful outcomes.

What AI Integration Could Look Like

If we want AI to matter in schools—not just appear in them—here’s what needs to happen first:

1. Create real production environments, not simulations — Students should use AI to build products for real audiences: local business tools, community resources, data analysis for local governments.
2. Make AI a teacher tool before a student curriculum — Give teachers AI assistants for lesson planning or feedback generation for six months before asking them to teach students about it.
3. Cut teacher administrative work by 50% first — Declare a one-year moratorium on new initiatives and use that time to eliminate pointless paperwork and meetings. Only then introduce AI tools.
4. Integrate, don’t segregate — No “AI class”—learning isolated subjects is already a broken premise. Students should use AI as a thinking partner embedded in real work: pressure-testing business models, mapping disease transmission for public health campaigns, or generating ad iterations for ventures they’re launching. It’s not a topic on a schedule—it’s a co-pilot in meaningful, multidisciplinary work.

The Core Problem

AI won’t fix school because school doesn’t reward the things AI is good at enhancing: curiosity, building, iteration, real-world problem-solving.

Until we rebuild the incentives and culture of school—until we stop measuring learning by what fits on a standardized test—AI won’t be transformative. 

It’ll be ornamental.

Worse, it will become another burden to teachers who are already asked to be instructors, therapists, parents, tech troubleshooters, data clerks, and disciplinarians.

Final Thought

AI could transform education, but not through executive orders. 

It needs something much rarer in education: patience, focus, and the courage to stop doing things that don’t work.

Great tools don’t fix broken systems—they just reveal the breaks more clearly.

If we don’t fix the foundation, AI won’t rescue schools. It will just reveal how irreparably cracked they’ve become.