The short answer: what a 2026 school AI policy covers
A school AI use policy in 2026 sets out who may use which AI tools, for what, and under what safeguards. The policies we reviewed for this guide share five elements: an approved-tools list with a vetting process, grade-band rules for students, rules for staff that keep a human responsible for decisions about students, privacy rules that keep personal information out of unapproved tools, and an academic integrity standard that says when and how AI use must be disclosed.
The policy environment has firmed up quickly. According to Ballotpedia's tracker of state AI guidance, education departments in 35 states had issued K-12 AI guidance as of July 2026. Ohio law required every district, community school and STEM school to adopt an AI policy by 1 July 2026, and Tennessee has required local boards to adopt one since the 2024-25 school year. Facts in this guide were checked on 8 November 2026. It is general information, not legal advice.
What federal guidance says
There is no federal law requiring schools to have an AI policy. The U.S. Department of Education has instead issued guidance on how federal funds may be used for AI and on the principles it expects schools to follow.
- July 2025 Dear Colleague letter. On 22 July 2025, the Department sent guidance on using federal grant funds for AI. It confirms that responsible AI uses can be allowable under existing programmes and stresses engaging parents and teachers in adoption decisions.
- Five principles. The Department summarises that letter as five principles for responsible AI: education technologies should be educator-led, ethical, accessible, transparent and protective of student data.
- August 2026 letter on education technology. On 20 August 2026, the Department issued a further Dear Colleague letter on technology and screen time, asking schools to choose tools for instructional value and to be transparent about how and when technology is used.
Existing privacy laws still apply to every AI tool: FERPA for education records, COPPA for online services used by children under 13, and the Children's Internet Protection Act for schools that receive E-rate funding. Our COPPA checklist for K-12 schools covers the vendor questions in detail.
What state guidance and laws require
Most state guidance is advisory, but a growing number of states have turned it into a legal duty. The table shows two binding examples alongside the broader picture.
| State or source | Instrument | What it requires |
|---|---|---|
| Ohio | Ohio Revised Code 3301.24, added by House Bill 96 | The Department of Education and Workforce had to publish a model AI policy by 31 December 2025; every district, community school and STEM school had to adopt an AI policy by 1 July 2026 and may adopt the model |
| Tennessee | Tennessee Code 49-2-203(a)(16) | Local boards must adopt a policy on AI use by students, teachers and staff for instruction and assignments, implemented from 2024-25, and report compliance to the state each 1 July |
| 35 state education departments | Advisory guidance (Ballotpedia, July 2026) | Recurring themes: AI literacy, bias, privacy and security, equitable access, academic integrity and stakeholder engagement |
Ohio's model AI policy for districts is a useful starting point even outside Ohio because it was written to be adopted as is. Wherever you are, check your state education department's page before drafting: many states have updated their guidance in 2026, and several legislatures considered new AI-in-education bills this year.
How real districts are writing their policies
District policies vary widely in how much student use they allow. Three published examples show the range. Each is summarised from the district's own page as read on 8 November 2026.
New York City Public Schools: grade-band limits and no AI decisions about students
NYC Public Schools' guidance on AI and screen time for 2026-27 bans student-facing generative AI from pre-K through grade 8. High school students may use only approved, vetted programmes, with AI literacy modules and teacher supervision. Staff may use AI for planning and operational tasks once a tool passes the city's data privacy and security review, but AI may not be used for grading, behaviour monitoring, or placement, promotion and graduation decisions. Schools must notify families when students take part in an approved programme.
Independent School District 12, Minnesota: the Human-AI-Human model
ISD 12's Policy 625, adopted on 15 June 2026, treats AI as an augmenting tool rather than a decision-maker under a "Human-AI-Human" model: a person frames the task, AI assists, and a person with the right expertise verifies the output. Students may use district-approved tools only as directed by teachers and must disclose AI use as each assignment requires. The policy restricts AI in disciplinary, placement and behavioural assessment decisions, bars entering personal or confidential information into AI tools, and requires vendors to identify AI components embedded in the products they supply.
Manasquan School District, New Jersey: policy plus a published AI plan
Manasquan's board adopted a policy on acceptable use of generative AI that requires the district to develop and publish a district AI plan. Pairing a short board policy with a more detailed, regularly updated plan is a common pattern: the policy sets principles, and the plan carries the tool list and procedures that change every term.
The elements most policies share
Comparing state guidance and district policies, the same building blocks appear again and again.
| Element | What it decides | Typical wording to look for |
|---|---|---|
| Scope and definitions | Which tools count as AI, including AI features inside existing products | Defines generative AI, deepfakes and "district-approved tool" |
| Approved tools and vetting | Who approves a tool and on what criteria | Privacy and security review before any classroom use |
| Student use by grade band | What students may do at each age | Bans or teacher-directed use for younger students; supervised use for older ones |
| Staff use | What teachers and administrators may use AI for | Planning and admin tasks allowed; a human stays responsible for outputs |
| High-stakes decisions | Where AI is not allowed | No AI-made grading, discipline, placement or graduation decisions |
| Data and privacy | What may never go into an AI tool | No personal, confidential or sensitive information in unapproved tools |
| Academic integrity | When AI use must be disclosed | Cite or declare AI assistance as each assignment requires |
| Transparency to families | What parents are told and when | Annual notice and notice when students join AI programmes |
| Training and AI literacy | Who is trained, and in what | Staff training plus age-appropriate AI literacy for students |
| Review cycle | How often the policy is revisited | At least annually, given how fast tools change |
A 15-point checklist for drafting or updating your AI policy
Use this checklist to draft a new policy or audit an existing one. It assumes a board policy plus administrative procedures, but works for a single school too.
- Check your state's law and education department guidance, and note any adoption or reporting deadline.
- Form a small working group: an administrator, teachers from each phase, IT, a special education lead, and parent and student voices.
- Inventory the AI already in use, including AI features switched on inside tools you already licence.
- Define key terms in plain language, including what counts as a district-approved tool.
- Set a vetting process for new tools that covers FERPA, COPPA, state privacy law, security and whether student data trains models.
- Write grade-band rules for student use, with clear defaults for the youngest students.
- List permitted staff uses and require a human to review any AI output before it affects a student.
- Name the decisions AI must not make: grading, discipline, placement, promotion, graduation and similar.
- State what information may never be entered into an AI tool.
- Set an academic integrity standard: when AI is allowed on an assignment and how students disclose it.
- Plan accessibility: confirm approved tools work with assistive technology and that IEP and 504 accommodations take priority.
- Decide how families are told about AI use, approved tools and any opt-out.
- Schedule staff training and AI literacy for students before the policy takes effect.
- Assign an owner for the approved-tools list and incident reporting.
- Set a review date, at least annually, and record what changed and why.
Making the policy work day to day
A policy only works if staff can see which tools are approved, students know the rules for each assignment and families get the same message from every classroom. Three operational habits help.
- Keep one approved-tools list. Publish it where teachers already look and record the approval date and reviewer for each tool.
- Tie disclosure rules to the assignment. Teachers state the AI rule for each assignment, so integrity cases are judged against a written expectation rather than a guess.
- Use one channel for family notices. Send AI notices through the same system you use for other school announcements so there is a dated record of what each family received.
Borderset supports this side of the work rather than the AI tools themselves. Family announcements keep a record of notices sent, exam management keeps assessment results tied to the student record, and role-based access and audit trails show who changed a record. Borderset also includes AI feedback, and your policy should cover it like any other AI tool, with a teacher reviewing output before it reaches a student or family. To see how that fits your policy, you can book a demo.
Related reading on Borderset
For where AI genuinely saves time in the school office, read AI in K-12 school operations. For assessment-specific questions, see AI grading in K-12: when it helps and when it doesn't and AI proctoring for K-12.
If your policy work sits alongside a new phone policy, our guide to school cellphone bans by state covers the 2026 laws. For vendor contracts, see what the AFT, UFT and Anthropic AI safety and privacy agreement covers, and for statewide technology plans that now include AI guidelines, read about the Kansas instructional technology policy.
For a large district example of a pause plus an advisory group, read about NYC's Technology in Schools Coalition. To vet AI and other tools before they reach students, use our guide to building a school edtech review process.
Frequently asked questions
Do schools have to have an AI policy?
There is no federal requirement. Some states require one: Ohio law required every district, community school and STEM school to adopt an AI policy by 1 July 2026, and Tennessee has required local boards to adopt one since the 2024-25 school year. Elsewhere, state guidance is usually advisory.
How many states have AI guidance for K-12 schools?
According to Ballotpedia, education departments in 35 states had issued AI guidance for K-12 public schools as of July 2026.
What does the U.S. Department of Education say about AI in schools?
Its July 2025 Dear Colleague letter says responsible AI uses can be allowable under federal grant programmes and sets out five principles: technology should be educator-led, ethical, accessible, transparent and protective of student data. An August 2026 letter asks schools to choose technology for its instructional value.
Should a school AI policy ban student use of ChatGPT?
Policies differ. NYC Public Schools bans student-facing generative AI through grade 8 and allows only vetted programmes in high school, while other districts allow teacher-directed use of approved tools. Most policies set rules by grade band rather than a single blanket ban.
What should never be entered into an AI tool at school?
Most policies bar entering personally identifiable, confidential or sensitive information about students, staff or families into any AI tool that the district has not approved and reviewed for privacy and security.
How often should a school review its AI policy?
At least once a year. AI tools and state guidance change quickly, and several districts commit to regular reviews with recommendations to the board.