top of page

How to Write an AI Policy for a K-12 Private School (and What to Decide First)

Oct 5
7 min read

A parent emails to ask whether students are allowed to use ChatGPT on homework. A teacher brings you an essay that sounds nothing like the student who turned it in. A department head wants to know which AI tools the school has approved.


Three questions in one week, and the honest answer to all three may be the same: we haven't decided yet.


If that's your school, I'm going to be honest with you because I care: you're behind.


This post walks through how to write an AI policy for private schools in the order the decisions actually need to be made, so the policy answers those questions before they reach your inbox.


Most schools are writing this policy after the fact.


Students and teachers didn't wait for a policy. In a national RAND survey from the 2024 to 2025 school year, 54% of students and 53% of English, math, and science teachers said they used AI for school (Doss, Bozick, Schwartz, et al., 2025). Only 45% of principals reported having any school or district policy or guidance on AI, and only 34% of teachers said they had a policy on AI and academic integrity.


Two more numbers from the same study stay with me. Over 80% of students said their teachers had not explicitly taught them how to use AI for schoolwork, and half of students said they worry about being falsely accused of cheating with it. Students are using a tool nobody has taught them, under rules nobody has written down, and they're nervous about getting it wrong.


A policy fixes the "rules nobody has written down" part. It's also the one part fully in your hands.


First, know which rules apply to a private school.


Private schools start from a different place than public districts, and a lot of the AI guidance floating around assumes you're a public district. Three sets of rules are worth understanding before you write anything.


  • Plan for FERPA. FERPA protects student education records at schools that receive U.S. Department of Education program funds, and whether a private school is covered depends on its funding (U.S. Department of Education, Student Privacy Policy Office, n.d.). Either way, families expect their children's records to be protected to that standard, and building your AI policy around FERPA's protections gives every teacher one clear benchmark for student data.

  • COPPA matters for students under 13. COPPA governs the companies behind online tools used by children under 13. A school can consent on a parent's behalf only when the tool is used for the school's educational purposes and for no other commercial purpose, and the FTC suggests schools make those notices available to parents (Federal Trade Commission, n.d.).

  • The tools have their own age rules. OpenAI's terms, for example, require users to be at least 13, with a parent or guardian's permission under 18 (OpenAI, 2026). UNESCO recommends an age limit for independent conversations with generative AI (Miao & Holmes, 2023).


Two federal signals point the same way. The Department of Education's 2025 guidance on AI calls for attention to privacy and for engaging parents in decisions about new technology (U.S. Department of Education, 2025). This post is a starting point, and I'd recommend running the privacy pieces past your school's legal counsel before the policy goes out.


The five decisions, in order


Most policies I read start with a statement about AI and the future. I'd suggest starting with the decisions instead, because each one makes the next one easier. Here are five worth working through, in the order I'd take them.


AI policy for private schools checklist: five decisions in order, who decides, approved tools, what never goes in, what counts as cheating, and grade-level expectations

  1. Who decides. A small group tends to work well: a school leader, someone from curriculum, whoever handles technology, a teacher or two, and a parent. Arizona's state guidance recommends a cross-functional team and encourages schools to bring students and families in early (Arizona Institute for Education and the Economy, 2026).

  2. Which tools are approved. A short list of tools you've actually reviewed, with school accounts where possible, is a strong place to start. Arizona's guidance names an approved tool list as one of the three first moves, and cautions that schools should not assume a platform's privacy practices meet legal requirements (Arizona Institute for Education and the Economy, 2026). The good news is that several companies leading AI in education now offer education accounts at no cost. ChatGPT for Teachers is free for verified U.S. K-12 educators through June 2028 (OpenAI, 2025). The Gemini app is included free in every Google Workspace for Education edition, and Google says data in those accounts isn't human reviewed or used to train its models (Google, n.d.). Microsoft 365 Copilot Chat comes at no additional cost with a school Microsoft 365 account (Microsoft, n.d.-a), and students 13 to 17 can use it only after an administrator turns it on (Microsoft, n.d.-b). If your school already runs on Google or Microsoft, the account you'd approve may already be in place.

  3. What never goes into an AI tool. One plain rule is worth writing down: no student names, grades, health information, or other identifying details in any tool the school hasn't approved. This is the line teachers cross by accident, usually while trying to save time on a report card comment.

  4. What counts as cheating. Defining AI-assisted work in your academic integrity policy, along with what students should disclose, gives everyone the same starting point. RAND's first short-term recommendation is clarity for teachers and students on what counts as cheating with AI (Doss, Bozick, Schwartz, et al., 2025). It's also worth deciding early that a detection score alone won't settle a case; Arizona's guidance says detectors should never be the sole basis for that call (Arizona Institute for Education and the Economy, 2026).

  5. What each grade level may do. This is the part parents ask about most, so it gets its own section.


Grade-level expectations


Here's one starting point I'd suggest for a K-12 private school, to adjust for the tools you approve and their age rules:


  • K-2: Teachers use AI to plan and prepare. Students see it only in teacher-led demonstrations, with no student accounts.

  • Grades 3-5: Teacher-led use on approved tools, with the teacher at the keyboard or in a supervised class activity. Students start learning what AI is and where it makes mistakes.

  • Grades 6-8: Students use approved school accounts for specific assignments, with the AI expectation written on the assignment. For students under 13, that means tools covered by the school's COPPA consent.

  • Grades 9-12: Broader use on approved tools, with disclosure every time AI touches submitted work.


RAND makes a point I agree with: elementary school is where students form habits, so the younger grades are worth including even though they are the least likely to be using AI tools (Doss, Bozick, Schwartz, et al., 2025).


A policy sets the boundaries. The assignment is where students actually meet them.

Where the AI Usage Scale fits.


A policy can say "students may use AI with permission." It can't say what permission looks like on Tuesday's history assignment. That gap is why I built the AI Usage Scale for Assignments and Syllabi: four levels, from AI-Free to AI-Empowered, with one label on every assignment and a short statement students attach to their work.


The policy says AI use must be disclosed. The scale tells students how, assignment by assignment. Naming a shared scale in your policy gives every teacher the same vocabulary.


Will a written policy stop every student from misusing AI? Not at all. A policy with clear expectations does mean that when a student crosses a line, they crossed a line that was explained to them, and the conversation that follows is about learning instead of guessing.


Put it in the documents you already have.


You may not need a new binder. TeachAI recommends adding AI language to three policies most schools already keep: responsible use, academic integrity, and privacy (TeachAI, 2025). Arizona's guidance adds the student code of conduct (for things like deepfakes used to harass a classmate) and the staff and student handbooks (Arizona Institute for Education and the Economy, 2026).


It also helps to name an owner and a review date. The tools will change before the school year ends, and a policy with a review date on it is one people trust.



Questions school leaders ask


Do private schools have to follow FERPA?


It depends on funding. FERPA applies to schools that receive U.S. Department of Education program funds, and the Department notes that private and parochial K-12 schools generally don't. Even so, we recommend writing your AI policy to FERPA's standard: families expect their children's records to be protected, and it gives teachers one clear rule. Your legal counsel can confirm where your school stands.


Should we ban AI for younger students?


You can keep student accounts out of the lower grades and still teach what AI is. Teacher-led demonstrations let young students learn that AI can be wrong, which is a lesson worth starting early.


Do we need AI detection software?


A good policy doesn't have to depend on it. Detection tools are not fully reliable, so we don't recommend using them to decide cases. Clear expectations on each assignment and a conversation with the student do more.


How often should we update the policy?


Once a year is a sensible minimum. Keeping the approved tool list on its own page lets it change without a full policy revision.


Your next step


A policy works only when every teacher can explain it to students the same way. The Educator AI Essentials Workshop is two hours with your whole faculty on your school's approved tools, and it covers the guardrails your policy sets: protecting student data, checking AI output, and the AI Usage Scale for assignments. You can see it on our Faculty Professional Development page or book a free consultation and we'll talk through where your policy stands.


Which of the five decisions is your school still working through?


References


Arizona Institute for Education and the Economy. (2026). Generative artificial intelligence in K-12 education: Guidance for Arizona schools and school systems (Version 26.01). Northern Arizona University. https://in.nau.edu/college-education/wp-content/uploads/sites/359/2025/05/AZ-GenAI-Guidance-v26.01.pdf


Doss, C. J., Bozick, R., Schwartz, H. L., Chu, L., Rainey, L. R., Woo, A., Reich, J., & Dukes, J. (2025). AI use in schools is quickly increasing but guidance lags behind: Findings from the RAND survey panels. RAND. https://www.rand.org/pubs/research_reports/RRA4180-1.html


Federal Trade Commission. (n.d.). Complying with COPPA: Frequently asked questions. https://www.ftc.gov/business-guidance/resources/complying-coppa-frequently-asked-questions


Google. (n.d.). Quickstart guide to Gemini and Gemini Notebook for Education. Google Workspace Help. https://knowledge.workspace.google.com/admin/getting-started/editions/quickstart-guide-to-gemini-and-gemini-notebook-for-education


Miao, F., & Holmes, W. (2023). Guidance for generative AI in education and research. UNESCO. https://www.unesco.org/en/articles/guidance-generative-ai-education-and-research


Microsoft. (n.d.-a). Learn about Microsoft 365 Copilot in Education. https://www.microsoft.com/en-us/education/products/copilot-in-education


Microsoft. (n.d.-b). Enable Copilot Chat for learners. Microsoft Learn. https://learn.microsoft.com/en-us/microsoft-365/education/guide/1-reference/enable-copilot-chat-for-learners


OpenAI. (2025). A free version of ChatGPT built for teachers. https://openai.com/index/chatgpt-for-teachers/


OpenAI. (2026). Terms of use. https://openai.com/policies/terms-of-use/


TeachAI. (2025). Sample considerations for existing policies. AI Guidance for Schools Toolkit. https://www.teachai.org/toolkit-addendums


U.S. Department of Education. (2025, July 22). U.S. Department of Education issues guidance on artificial intelligence use in schools, proposes additional supplemental priority. https://www.ed.gov/about/news/press-release/us-department-of-education-issues-guidance-artificial-intelligence-use-schools-proposes-additional-supplemental-priority


U.S. Department of Education, Student Privacy Policy Office. (n.d.). To which educational agencies or institutions does FERPA apply? https://studentprivacy.ed.gov/faq/which-educational-agencies-or-institutions-does-ferpa-apply

bottom of page