Your AI Policy Is Not AI Literacy

Picture a school that just finished its AI policy. Twelve pages. Legal reviewed it. The board approved it. It covers acceptable use, academic integrity, data privacy, and a clear process for violations. By every administrative measure, it is done.

The policy is finished. The hard part has not started.

Now walk down the hall and ask three teachers a few questions. Ask them to explain the difference between an AI model that reasons through a problem and one that simply predicts the next likely word. Ask them how they would redesign a single assignment so that AI cannot shortcut it. Ask them what they actually say to a student who admits, halfway through a conference, that they used AI on the draft.

Most cannot answer. And that distance, between the policy sitting in the shared drive and the skill that is missing from the room, is the most important thing happening in education right now. Call it the Policy-Literacy Gap, because once you can name it, you start seeing it everywhere.

The policy surge is real, and it is fast

Institutions are not ignoring AI. They are governing it at remarkable speed. FutureEd is currently tracking 53 AI-in-education bills across 25 states, covering what students learn about AI, how schools integrate it, and what guardrails apply. The State University of New York adopted a single AI policy across all 64 of its campuses in one board vote. At the federal level, an executive order has elevated AI literacy as a national priority, the U.S. Department of Education has named AI a grantmaking priority, and AI literacy is scheduled to appear on the 2029 PISA exam for the first time.

If activity were the same thing as progress, education would be in excellent shape. But it is not the same thing. And the gap between the two is widening every week a new policy gets announced.

A policy is a document. Literacy is a capability.

Here is the embedded truth most institutions are working hard to avoid. A policy tells people what they are allowed to do. Literacy determines whether they can actually do it well. Those are two different projects, and finishing the first one does not produce the second.

A teacher can read every line of a twelve-page AI policy and still have no idea how to use AI to build a sharper rubric, recognize AI-assisted work, or coach a student through responsible use instead of punishing them for it. The policy governs a competence it quietly assumes already exists. The literacy that would make the policy meaningful is the part nobody put on a calendar.

Now, you might be thinking that a policy at least sets guardrails, and that guardrails are worth something. They are. But guardrails on a road that nobody has been taught to drive do not produce safe drivers. They produce a slower, better-documented crash.

The data shows exactly where the gap lives

This is not a hunch. The Center for Democracy and Technology found that 85 percent of teachers and 86 percent of students used AI during the last school year. That is effectively universal adoption. The same report found that fewer than half of them received any training or guidance from their institution.

Read those two numbers together, because together is the only way they tell the truth. Nearly everyone is using AI. Fewer than half were ever taught how. The policy may be new, but the usage is already old, and it has been running untrained this entire time.

The mistake has a name

When an institution writes the policy, announces the policy, and then treats the problem as solved, that is policy theater. It looks like leadership. It produces a document, a press line, and a satisfying sense of closure. What it does not produce is one teacher who is measurably more capable on Monday than they were on Friday.

Policy theater is seductive because it is finite. You can finish a policy. You cannot finish literacy, because literacy is a practice, not a deliverable. And that is exactly why so many institutions default to the document and skip the practice. One has an end date you can celebrate. The other asks for a habit you have to keep.

The institutions getting this right have stopped asking whether the policy is done and started asking what their people actually learned this month. If you want a closer look at what that pivot sounds like in practice, our piece on asking better questions about AI in the classroom walks through it.

What AI literacy actually looks like

So what is the alternative. Not a longer policy. A different unit of progress entirely.

Real AI literacy gets built the way every durable skill gets built, through small, repeatable wins that compound over time. That is the whole idea behind SeedStacking, the Harvest Kernel methodology, and it moves through four phases. Seed is first contact, a five to fifteen minute task that produces one concrete result. Sprout is repetition with slight variation, until the tool stops feeling foreign. Grow is applying the skill to real work that genuinely matters to you. Harvest is producing something you can share, teach, or hand to a colleague.

Notice what that structure does to the problem. It turns becoming AI literate from an overwhelming institutional mandate into a sequence of fifteen-minute tasks. A faculty member does not need a semester-long course or another policy revision before they can start. They need one good Seed this week. A rubric drafted with AI and then sharpened by hand. One assignment rebuilt so it rewards thinking the AI cannot fake. One honest conversation with a student, rehearsed first. Then another Seed next week.

That is not a lowering of the bar. It is the only way anyone has ever cleared it.

The institutions that will be fine

Here is the contrast that should guide every leadership decision in the next year. The schools that thrive will not be the ones with the longest or most carefully worded AI policy. They will be the ones whose teachers can think clearly with AI, and just as clearly without it.

A policy ages the moment a new model ships. A literate, confident educator adapts on their own, because literacy is portable in a way that a document never is. One is a thing you have to defend. The other is a capability that defends itself.

When the federal government made AI literacy a funding priority earlier this month, it signaled the same thing the latest coverage is signaling now. The conversation is moving past whether to govern AI and toward whether anyone can actually use it well. The policy era is ending. The literacy era is the one worth preparing for.

If you are the person responsible for your institution’s AI strategy, the honest question is not whether the policy is finished. It is what your people actually learned this month. If you do not have an answer to that, you do not yet have a strategy. You have paperwork.

The Policy-Literacy Gap does not close with another meeting or another document. It closes one Seed at a time, one educator at a time, one small win that quietly compounds into confidence. That work is not glamorous and it does not produce a press release. It just produces teachers who are ready. Which, in the end, was the entire point.