What Should a School AI Policy Include? A Checklist for Administrators

Students are already using AI chatbots for homework, essays and study help. Teachers are using AI to draft lesson plans, rubrics and parent emails. And in many schools, nobody has written down what’s allowed. The result is confusion: one teacher bans AI completely, the next encourages it, and students and parents aren’t sure where the lines are.

A clear school AI policy fixes that. It gives teachers backing, gives students fair rules and protects student data. This checklist covers what to include, whether you’re a principal drafting a building policy or a district leader updating board policy.

This is general information, not legal advice. Work with your district’s legal counsel on privacy and compliance language.

Why every school needs an AI policy now

Without a policy, schools face real risks:

• Inconsistent discipline. Students get punished for AI use in one class that’s encouraged in another.
• Privacy exposure. Staff paste student names, grades or IEP details into consumer AI tools.
• Unreliable detection. Teachers rely on AI detectors that can produce false positives. Our guide on how teachers check for AI writing explains the limits.
• Missed opportunities. Teachers who could save hours with AI hold back because they’re unsure what’s allowed.

Federal leaders are paying attention too. In July 2025, the US Department of Education issued a Dear Colleague letter explaining that federal grant funds can support responsible AI uses, such as AI-based instructional materials, AI-enhanced tutoring and college and career advising. It set out principles including educator-led use, ethics, accessibility, transparency and data protection, and said AI systems must comply with federal privacy laws such as FERPA.

The school AI policy checklist

Use this as your outline. Each item is covered in more detail below.

Policy section	Key question it answers	Owner
Purpose and scope	Who and what does this policy cover?	District leadership
Allowed and prohibited uses	What can students and staff use AI for?	Curriculum and instruction
Assignment-level AI levels	How do teachers signal AI rules per assignment?	Teachers
Disclosure and citation	How do students show where AI helped?	Teachers, librarians
Academic integrity process	What happens when misuse is suspected?	Administrators
Student data privacy	What data can go into AI tools?	Privacy officer, IT
Approved tools and vetting	Which tools are approved, and how are new ones reviewed?	IT, procurement
Staff use	How can teachers and staff use AI?	HR, leadership
Training	Who gets trained, and on what?	Professional development
Communication with families	How are parents informed?	Communications
Review cycle	When is the policy updated?	Policy committee

Scope and allowed uses

1. Purpose, scope and definitions

Start with a short statement of why the policy exists: to support learning, protect students and set fair expectations.

Then define:

• What counts as AI for this policy, including chatbots, writing assistants, image generators and AI features built into tools you already use.
• Who it covers: students, teachers, staff, contractors and volunteers.
• Where it applies: school devices, school accounts and schoolwork done at home.

Keep definitions simple. Families should understand them without a technical background.

2. Allowed and prohibited uses

Rather than a blanket ban or blanket approval, spell out uses by category.

Usually allowed, with teacher permission:

• Brainstorming ideas and outlines
• Explaining a concept in a different way
• Practice quizzes and study guides
• Feedback on grammar and clarity
• Translation support for multilingual learners
• Accessibility support, such as text-to-speech or reading level adjustments

Usually prohibited:

• Submitting AI-generated work as your own without disclosure
• Using AI during tests unless explicitly allowed
• Creating images or audio of real people without consent, including deepfakes of classmates or staff
• Entering other people’s personal information into AI tools
• Using AI to harass, bully or cheat

Our guide to how students can use AI for homework shows what responsible use looks like from the student side.

Use assignment-level AI levels.

Many schools use a simple scale so each assignment states its AI rules clearly:

Level	Meaning	Example
No AI	AI tools not allowed	In-class essay, test
AI for planning	Brainstorming and outlining only	Research paper topic selection
AI with disclosure	AI allowed for drafting or feedback, with a note on how it was used	Revising a draft
AI encouraged	AI is part of the task	Critiquing an AI-written answer

Putting the level on every assignment removes guesswork and makes enforcement fairer.

Key takeaway: The most useful AI policies don’t just say yes or no. They give teachers a shared vocabulary for assignment-level rules and give students a clear way to disclose AI help.

Disclosure and academic integrity

3. Disclosure and citation

Students should know exactly how to show AI help. Options include:

• A short AI use statement at the end of an assignment: which tool, what it was used for and what the student changed.
• Citing AI output using the style guide your school already teaches, such as MLA or APA, which both publish guidance on citing generative AI.
• Keeping drafts or chat histories when asked, so teachers can see the process.

Disclosure should be normal, not shameful. If students fear punishment for being honest, they’ll hide AI use instead.

4. Academic integrity and handling suspected misuse

Spell out a fair process:

1. Teacher reviews the work against the student’s past work and the assignment’s AI level.
2. Conversation first. Ask the student to explain their process, ideas or sources.
3. Don’t rely on AI detectors alone. Detection tools can be wrong, and policies should say a detector score is never enough evidence by itself.
4. Follow existing integrity procedures for consequences, with escalation steps and parent contact.
5. Focus on learning, such as redoing the assignment, especially for first incidents.

Also consider designing assignments that make misuse less useful: in-class writing, oral explanations, personal reflection and process checkpoints.

Student privacy and approved tools

5. Student data privacy

This is the section most likely to have legal consequences. Cover:

• FERPA. Student education records are protected. Staff shouldn’t enter identifiable student information, grades, disability information or discipline records into AI tools that aren’t approved and covered by a data agreement.
• COPPA. For students under 13, the FTC’s COPPA guidance says schools can consent on behalf of parents to online services used solely for educational purposes, but not when the operator uses the data for its own commercial purposes, such as behavioral advertising. The FTC amended the COPPA Rule in April 2025.
• State student privacy laws. Many states have their own student data privacy laws. Our guide to US state data privacy laws and state AI laws can help you find relevant rules.
• Data agreements. Require a signed data privacy agreement before any AI tool processes student data.
• Data minimization. Use de-identified examples when testing tools, and turn off settings that let vendors train models on school data where possible.

6. Approved tools and vetting

Maintain a list of approved AI tools with the purpose and age range for each. Before approving a tool, check:

• [ ] Age requirements. Many consumer AI tools require users to be at least 13. OpenAI’s terms, for example, set a minimum age of 13 and require parent or guardian permission for users under 18.
• [ ] Privacy terms. Does the vendor use student data for training or advertising?
• [ ] Data storage and deletion. Where is data stored, and can the school request deletion?
• [ ] Admin controls. Can IT manage accounts, settings and access?
• [ ] Accessibility. Does the tool work with assistive technology?
• [ ] Accuracy and bias. Has the tool been tested for errors and AI hallucinations in your subject areas?
• [ ] Cost and contract terms. Including what happens when a free pilot ends.

Give teachers a simple way to request review of new tools, so they aren’t tempted to use unapproved ones.

Staff use and training

7. Staff use of AI

Teachers and staff need rules too:

• Allowed: drafting lesson plans, rubrics, newsletters and differentiated materials, with human review. See how teachers use AI to plan lessons.
• Allowed with care: feedback on student writing, only in approved tools and without identifiable data unless a data agreement covers it.
• Not allowed: making final grading, placement or discipline decisions based solely on AI output.
• Transparency: tell students and families when AI is used in ways that affect them.

If your district already has a broader AI acceptable use policy for employees, align the two documents.

8. Training for staff, students and families

A policy without training rarely changes behavior. Plan:

• Staff training on the policy, approved tools, privacy rules and assignment design.
• Student lessons on AI literacy: how AI works, its errors, bias, citation and ethics.
• Family information sessions or guides explaining the rules and how parents can support learning at home.

The Department of Education’s July 2025 letter lists training educators, providers and families on responsible AI use as an allowable use of federal funds.

Where to start: build your policy in 6 steps

1. Form a small working group with teachers, an administrator, IT, a privacy lead, a librarian, students and parents.
2. Survey current use. Find out how staff and students already use AI.
3. Draft using the checklist above, borrowing from your state education agency’s guidance if it has issued any.
4. Pilot with a few teachers for a grading period.
5. Train and communicate before full rollout.
6. Review at least once a year, because tools and laws change quickly.

For more context on where classroom AI is heading, see our overview of AI in education.

Conclusion

A good school AI policy is practical, not perfect. It defines allowed uses, gives teachers a shared way to set assignment-level rules, makes disclosure normal, protects student data and invests in training. Start with a clear first version, pilot it and plan to update it every year. Students are already using AI. The policy decides whether they use it to learn or to cut corners.

Sources

• US Department of Education: Guidance on AI use in schools (press release) — July 2025 guidance and proposed AI grant priority
• US Department of Education: Dear Colleague letter on federal grant funds and AI — allowable uses and responsible AI principles, including FERPA compliance
• FTC: Complying with COPPA, Frequently Asked Questions — school consent for educational services and 2025 rule amendment
• OpenAI: Terms of Use — minimum age and parental permission requirements