This guide is written for superintendents, curriculum directors, and IT leads who have been asked to produce a district AI policy and do not have a full-time policy specialist on staff. It covers the sections every K-12 AI policy needs, the legal anchors to cite, the grade-band decisions that actually matter, and the common mistakes that send drafts back for rewrite.

Why districts need a policy now

Teachers are making AI decisions every day — whether the district has a policy or not. Without written guidance, two teachers in the same building apply different rules to the same assignment, and the first time a student or parent escalates, the district is making policy under pressure. A written policy removes that pressure.

The regulatory environment is also moving. Texas adopted TEC §11.169 in 2024, requiring districts to have an AI policy. Other states are following. FERPA still governs how student records are handled, and AI tools that ingest student data fall under those same rules. Most districts are finding that the question is not whether to write a policy, but how quickly they can produce one that holds up.

What goes in a K-12 AI policy

A defensible policy has eleven sections. The first nine form the policy itself (what the board adopts). The last two — teacher and student guidelines — translate the policy into plain language for the people who have to live with it.

• Purpose — why the district is adopting this policy and what it aims to protect.
• Scope — who the policy applies to (staff, students, contractors) and where (classroom, home, extracurriculars).
• Student Use — allowed uses, by grade band. This is the heart of the policy.
• Prohibited Uses — specific categories of behavior that are never permitted.
• Teacher Responsibilities — what staff must do before, during, and after AI-related assignments.
• Lesson Planning — how teachers should incorporate AI when the policy allows it, and how to design around it when it doesn't.
• Privacy — the FERPA floor plus any district-specific rules on student data, IEPs, and behavior records.
• Transparency — when AI use must be disclosed (on student work, in communications with parents, in district-facing materials).
• Enforcement — how violations are handled, consistent with the district's existing discipline framework.
• Teacher Guidelines — practical staff-facing rules derived from the policy.
• Student Guidelines — grade-appropriate rules written in student language.

Legal and regulatory anchors

A K-12 AI policy should cite its sources, not because citation makes it feel official, but because the district will get questions from counsel and the board about the basis for each rule. The major anchors are:

• FERPA — governs any handling of personally identifiable student records, including records that pass through AI tools. Names, IEPs, grades, and behavior records fall under FERPA regardless of whether the tool is an AI.
• COPPA — applies to collection of personal information from students under 13. Relevant when AI tools are used with elementary students.
• TEC §11.169 — requires Texas districts to adopt an AI policy covering student use.
• TEA AI Guidance — advisory Texas state-level guidance; the best reference for aligning a Texas district policy with state expectations.
• NIST AI Risk Management Framework — voluntary federal framework; useful language for the Enforcement and Privacy sections.
• TEKS — Texas learning standards. Relevant for deciding which AI uses are compatible with the skills students are expected to demonstrate.

Grade-band gradient

The single most common mistake in K-12 AI policy drafts is treating "student use" as a single decision across all grade bands. A defensible policy names the gradient explicitly.

K-5 (Elementary): most districts restrict student AI use to teacher-directed activities. Students do not interact with AI tools independently. Rationale: developmental appropriateness, parental consent complexity, and COPPA.

6-8 (Intermediate): limited independent use on specific teacher-assigned tasks. Disclosure is typically required on submitted work. Academic integrity rules start to bite here.

9-12 (High): broader independent use, with the teacher setting boundaries per assignment. Disclosure policy varies (always / when substantial / teacher discretion). AI skills start showing up in college and career expectations, which argues for deliberate exposure.

Enforcement and academic integrity

Enforcement language should not invent a new discipline framework. It should route violations into whatever the district already uses — progressive discipline, restorative practices, or zero-tolerance for major incidents. The policy's job is to name which behaviors count as violations; the existing framework handles what happens next.

Two enforcement points are worth calling out explicitly. First, AI-detection tools (GPTZero, Turnitin's AI flag, etc.) have high false-positive rates and should not be the sole basis for an academic integrity finding. Second, suspected AI misuse should be handled through a conference and a look at the student's drafting history before consequences are applied — consistent with how other integrity issues are handled.

Getting it adopted by the board

Boards adopt what they can defend publicly. Three things make adoption easier:

• A policy that cites its legal anchors (FERPA, applicable state code, TEA guidance) removes the "what's the basis for this rule?" question before it is asked.
• A grade-band gradient lets board members see themselves answer the parent question "but what about my fifth-grader vs my high schooler?" without having to read the whole document.
• Teacher and student guidelines attached to the policy signal that adoption is not the end — the district has already thought about implementation.

Most districts adopt AI policy alongside a lighter-weight classroom rollout plan. The policy is the legal document; the rollout plan is the operational one.

Common mistakes to avoid

The drafts that come back for rewrite usually share one of these four problems.

• Naming specific tools. "Students may not use ChatGPT" ages badly and has to be amended every time a new tool launches. Name categories of behavior instead.
• Pretending the choice is binary. A single allowed/prohibited line across K-12 is not credible. The gradient matters.
• Ignoring teacher use. Teacher use of AI for lesson planning and feedback is a real question. Policies that only address student use leave teachers operating in the gray.
• No disclosure standard. Without a named rule for when AI use must be disclosed, classroom enforcement becomes per-teacher, which is exactly what the policy is supposed to prevent.

If you want a starting draft that has the right structure and the right anchors built in, Jombli generates one for you in under 15 minutes.

Adoption checklist

Ten concrete steps a district can take this month to move from “we should have an AI policy” to “we have an adopted AI policy.”

1. Confirm the decision-maker. In most districts, that is the superintendent; in some, a cabinet subcommittee.
2. Inventory AI tools currently in use. Count the gap between “approved” and “in use.” That gap is the reason for the policy.
3. Draft the policy from a framework. Use this guide or a generator. Do not start from a blank page.
4. Route to counsel before first reading. Give counsel at least two weeks.
5. First reading at a board meeting. Invite public comment.
6. Adopt at the second board meeting after incorporating comment.
7. Publish on the district website alongside the acceptable-use policy and student code of conduct.
8. Brief principals. They operationalize the policy with teachers.
9. Distribute teacher guidelines and the teacher quick-reference. These are the staff-facing versions.
10. Schedule the first annual review. Mark it on the board calendar so it actually happens.