Orbrya
All posts
Orbrya2026-03-12

School Ai Policy Gap Families

7 min read Here's a number worth sitting with for a moment. 96%. That's the share of families with elementary-aged children who either didn't know about any school AI policy or said their school hadn

Here's a number worth sitting with for a moment.

96%.

That's the share of families with elementary-aged children who either didn't know about any school AI policy or said their school hadn't communicated one, according to a nationally representative CRPE and USC survey conducted in spring 2025.

For secondary school families, the figure is 83%.

This isn't a fringe finding from a small sample. It's a probability-based national survey, the kind with the methodology to actually generalize. And what it's telling you is that the overwhelming majority of American families are navigating one of the most significant shifts in how children learn and work, the arrival of capable AI tools in every pocket and on every desk, almost entirely without institutional guidance.

That's not a reason to panic. It's a reason to act. Because the families who understand this gap and fill it themselves aren't just keeping up. They're getting ahead.

What the Policy Landscape Actually Looks Like

The narrative you might have heard, that schools are aggressively banning AI, doesn't match the data.

A 2024 EdWeek Research Center survey of 924 educators found that only about 7% of districts have banned AI for both teachers and students. Another roughly 20% have banned it for students only. The vast majority of schools haven't banned anything. They've done something that's arguably more problematic: they've said nothing coherent at all.

A 2025 RAND survey of principals found that only 45% reported having any school or district AI policy. The Gallup/Walton Family Foundation survey found that only 19% of teachers say their school has a meaningful AI policy, meaning the principal might think a policy exists, but the people in classrooms with students either don't know about it or don't find it useful.

Among the schools that do have policies, most focus on a single question: is using AI considered cheating? That's an academic integrity question, not an AI literacy question. It tells students when they're not allowed to use AI. It tells them nothing about how to evaluate AI when they do use it, which they will, regardless of policy. College Board data found 84% of high school students used AI for schoolwork in 2025. The tools are in use.

The gap isn't between permissive and restrictive schools. It's between schools that are teaching AI literacy and schools that aren't. Almost none of them are.

Why Schools Are Moving Slowly

Why do most schools have no AI policy? The short answer is that the pace of institutional change in education is structurally slow, and the pace of AI development has been structurally fast. School systems operate on multi-year curriculum cycles, require broad stakeholder consensus before adopting new policies, and depend on teacher training infrastructure that doesn't yet exist for AI. The result is a policy gap that is real, measurable, and unlikely to close quickly.

The longer answer involves what teachers are actually working with. A 2025 Gallup/Walton survey found that 68% of teachers received no AI training during the 2024-25 school year. The CDT found that only 28% of teachers received guidance on how to respond if they suspected a student used AI inappropriately. And RAND's nationally representative survey found that only 35% of district leaders report providing any AI training for students at all, while explicitly noting that elementary schools are being largely skipped in AI guidance rollouts despite the fact that nearly half of elementary teachers are already experimenting with AI tools on their own.

Teachers aren't ignoring AI because they don't care. They're navigating it without the training, the frameworks, or the institutional support to do it well. The guidance that does exist tends to arrive from district administrators who are themselves working without clear direction from state or federal policy, which is still being written.

California's AB 2876, signed in October 2024, directs the state to incorporate AI literacy into curriculum frameworks at the next scheduled revision, which could be years away. The federal AI education executive order signed in April 2025 established a task force and directed agencies to develop resources, but included no new funding. These are early signals that the system is moving, not evidence that it has arrived.

The schools that will teach AI literacy well are probably two to four years away from doing it consistently. That's a realistic read of how curriculum development, teacher training, and policy implementation actually work, not pessimism.

What the Gap Means for Your Child Right Now

Two to four years is a long time in a child's education. A student who is 10 today will be 12 to 14 before their school has a coherent AI literacy approach, assuming it develops on the optimistic end of that timeline.

That window matters because the habits being formed now are the habits that will persist. A RAND survey found that 54% of students were already using AI for schoolwork in 2025. The usage habits are developing whether or not the evaluative skills are developing alongside them. As we covered in our post on the Brookings doom loop research, the combination of high usage and low evaluative skill is precisely where the cognitive risk concentrates.

The families who act in this window aren't jumping ahead of some future requirement. They're filling a present gap, one that the data shows is real, widespread, and not going to close on its own anytime soon.

The Practical Advantage of Moving First

In any skill domain, the people who develop genuine competence early tend to compound that advantage over time. Not because they're inherently smarter, but because competence creates confidence, confidence creates practice, and practice creates deeper competence.

AI literacy is no different. A child who develops strong verification habits now, who learns to treat AI as a tool to direct and evaluate rather than a source to trust, will have been practicing those skills for years by the time their peers encounter them in a formal classroom setting. That gap doesn't disappear when the school finally catches up. It becomes part of how that child thinks.

This is the same advantage that families who introduced reading early, or mathematical reasoning early, or scientific thinking early have always had. The school eventually teaches the skill. But the child who already has it learns at a different level.

What You Can Do That Schools Currently Can't

The institutional constraints that slow schools down don't apply to you.

Schools have to move at the pace of curriculum committees, district approvals, teacher training cycles, and state standards. They have to reach 30 students simultaneously with wildly different starting points. They have to navigate parental concerns across the full spectrum from "AI is ruining education" to "why aren't we teaching more of this."

You have one child, or a few. You know them. You can meet them where they are, adjust in real time, and start tonight without waiting for anyone's approval.

The activities we've described in this series so far, the Spot the Lie game and the "How do you know?" habit, require no curriculum, no technology, and no expertise in AI. They're conversation patterns and thinking habits that transfer directly to AI use because they build the underlying skill: the reflex to verify before accepting.

That's what schools haven't figured out how to teach consistently. And it's what any family can start building this week.

A Word for Families in Traditional Schools

If your child attends a traditional school, the policy gap is your gap too, maybe more directly than for families who are already providing their child's full education at home.

Your child is using AI. The 84% figure from College Board isn't an abstraction. And if their school is among the 55% with no formal AI policy, they're doing it without any institutional framework for what good AI use looks like.

That doesn't mean you need to replace what the school is doing. It means adding what the school isn't doing. Fifteen minutes a week. One question at dinner. The occasional session working through an AI response together, asking what's right, what might be wrong, and how you'd check.

The school will eventually provide this. But eventually isn't now. And the families who treat this as their responsibility rather than waiting for the institution to catch up are the ones whose children arrive at "eventually" already prepared.

The Upside of the Gap

It's worth ending on this because it's genuinely true: the policy vacuum that feels like a problem is also an opportunity that won't exist forever.

Right now, AI literacy is not being systematically taught. The families who teach it have a real head start over families who don't. The children who develop verification habits now will have years of practice before those habits become standard curriculum.

That window closes as schools figure this out. The advantage is available precisely because the gap exists.

You're reading this at the right moment. The question is just what you do with it.

Next week we go deeper into the research on why verification specifically, not just AI usage, is the skill worth building. We'll cover the two types of critical thinking identified by researchers at King's College London, and why only one of them predicts the outcomes that matter.

Sources cited in this post: