School Starts in Six Weeks and Our District Still Has No AI Policy. Where Do I Start?

It is late July. The board is asking what the district's AI position is, a parent has already emailed about it, three teachers have quietly built their whole workflow on a chatbot, and the policy binder contains nothing newer than the device agreement. If "our district has no AI policy and school starts soon, where do I start" is roughly what you typed into a chatbot before landing here, this is the checklist you wanted.

School starts in six weeks and we have no AI policy. Where do I start?

Start with a one-page interim policy that makes four decisions, names one accountable owner, and explicitly says a full policy follows this year. Not a committee, not a 40-page framework borrowed from a bigger district, and not silence. The interim page does two jobs: it gives every principal the same answer for week-one questions, and it buys your real policy process the time to be done well. Perfection in October is worth less than clarity in August, because the school year does not wait for consensus.

What must an interim AI policy actually cover?

Four things: approved tools, protected data, disclosure, and staff use. First, name the specific AI tools staff and students may use, even if the initial list is one item long. Second, define what may never be entered into any AI tool: student names, IDs, grades, IEP and health information, anything identifiable. Third, set the disclosure norm: teachers tell students when AI rules apply to an assignment, students disclose when they used it. Fourth, state what staff may use AI for (drafting, planning, communication prep) and what stays human (grades, discipline decisions, anything involving identifiable student records). Every hard question you get in September maps back to one of those four lines.

Teachers are already using AI without asking. How do I get ahead of it?

Legitimize the honest lane instead of hunting the quiet users. Survey staff anonymously in August about what they already use AI for: the results will be higher than the board expects and useful for the real policy later. Then tell staff plainly: here is the interim policy, here is what is now explicitly fine, here is the one line you may not cross (protected student data), and no one gets punished for what they did before the page existed. You convert your early adopters into your pilot group, which is cheaper than any consultant and more persuasive with skeptical staff than any mandate.

Want the complete system? The interim one-pager as an editable template, the platform evaluation scorecards, the staff survey, and the rollout calendar are all in The District AI Playbook (Book 2). → Get The District AI Playbook ($49.99)

What about FERPA and student data?

Treat every general-purpose AI tool as a non-approved vendor until it has a signed data privacy agreement, because that is what it is. FERPA compliance in the AI era is less about the technology and more about the same vendor discipline districts already apply to gradebooks and learning platforms: know where data flows, have agreements in place, and keep identifiable records out of tools that lack them. The safe interim stance fits in one sentence: no personally identifiable student information in any AI tool, period, until the district signs an agreement that covers it. Mapping those data flows building by building is the entire subject of our companion volume Student Data in the Agent Era (Book 3).

Do we ban AI for students this year or allow it?

Neither extreme: allow it in labeled contexts, prohibit it in labeled contexts, and make teachers the labelers. A district-wide ban collapses on contact with take-home assignments, and a district-wide green light terrifies your veteran English department. The workable middle is assignment-level labeling: teachers mark work as AI-allowed, AI-assisted with disclosure, or AI-free, using consistent district language. Students get clarity, teachers keep authority over their own pedagogy, and the district has one vocabulary when a dispute reaches a principal's desk. For the classroom-level view of what agentic AI tools actually do, The Agentic Classroom (Book 1) is the primer to hand a curious principal.

Want the complete system?

The full playbook for administrators, the interim policy template, platform comparisons for the major education AI offerings, budget and PD planning, board presentation slides, and the January transition plan to a permanent policy, is in The District AI Playbook, Book 2: Policy, Platforms and Rollout for the Deadline Era. Written for the people who have to answer the email, not admire the problem. Through August 31, checkout code AISCHOOL20 takes 20 percent off.

→ Get The District AI Playbook ($49.99)

Frequently asked questions

Do we need board approval for an interim AI policy?
Usually not for an administrative guideline, which is what a one-page interim policy is. Check your local governance rules, brief the board rather than surprising them, and reserve formal adoption for the permanent policy.

Should we just ban AI until the full policy is ready?
A temporary ban is defensible only if you enforce it, and no district can enforce it on take-home work. An interim policy with a hard line on student data gives you more actual control than a ban that everyone knows is theater.

What is the biggest mistake districts make with AI policy?
Writing policy about the technology instead of about decisions. Tools change quarterly. A policy built on who approves tools, what data is protected, and how use is disclosed survives every product cycle.

Does FERPA already cover AI tools?
FERPA covers education records wherever they go, including into an AI tool, and that is precisely the problem: pasting identifiable student information into an unapproved tool can create a disclosure. The safe default is no identifiable data in any AI tool without a signed agreement.