AI Guidance for Teachers: What to Do When Your School Has None

Most teachers using AI in 2026 are doing it without any official AI guidance for teachers from their school or district.

According to a May 2026 Gallup and Walton Family Foundation survey of more than 2,000 U.S. public school teachers, 82% said they have received no formal guidance on how to apply AI tools to their work across multiple task areas. No more than 9% of teachers reported receiving any written policy for any specific AI-related task — including lesson planning, grading, feedback, or one-on-one instruction.

The gap isn’t because schools aren’t aware of the problem. It’s because the technology is moving faster than institutions can respond. A RAND 2025 report confirmed the same pattern: while AI use by teachers jumped more than 15 percentage points in a single year, fewer than half of principals reported having any AI use policy at their school.

This leaves most teachers in an awkward position — using a powerful tool daily, without clear institutional expectations about what’s appropriate, what’s risky, and what they’re accountable for if something goes wrong.

This article gives you a practical framework for building your own AI guidance when your school hasn’t provided one.

Why the Absence of AI Guidance for Teachers Matters More Than You Think

The most obvious risk of operating without guidance is making a mistake you didn’t know was a mistake — sharing student data with an unapproved tool, using AI output without review in a way that misleads a parent, or creating materials that violate academic integrity norms in ways your district hasn’t defined yet.

But there’s a subtler cost that the Gallup research identified: teachers who lack clear expectations — not just about AI, but about teaching generally — report higher burnout and lower job satisfaction than those with clear, realistic guidance. Ambiguity about what’s expected of you is its own stressor, separate from the actual workload.

The school principal quoted in the EdWeek coverage of the Gallup report put it well: when clear structures are in place, they liberate the user rather than restrict them. Without those structures, every decision about whether to use AI, and how, becomes a micro-judgment call that costs mental energy — exactly the resource burned-out teachers have the least of.

Building your own framework doesn’t replace institutional guidance. But it does give you a consistent set of personal rules that reduce the decision fatigue of navigating AI use one task at a time.

Your Four-Part Personal AI Guidance Framework

The framework below is organized around the four questions that matter most for a teacher using AI without school policy. Work through them once, write down your answers, and you have a personal policy you can follow consistently.

Part 1: Define What You Will and Won’t Input

The most consequential decision in AI use is what information you allow into the tool. Consumer-grade AI tools — the free tiers of ChatGPT, Gemini, and similar platforms — are not covered by district Data Processing Agreements. Anything you type into them could be retained, used in training, or accessed outside your control.

Your personal rule for inputs should be simple and binary:

Allowed: anything that doesn’t identify a specific student, parent, or colleague. Your own professional knowledge, generic scenarios, curriculum content, subject matter you’re explaining, communication templates without names, planning frameworks.

Not allowed: student names combined with any other identifying detail, grades or scores attached to identifiable students, IEP or 504 content, disciplinary notes, parent contact information, or photographs of students or their work.

If you want the full breakdown of what FERPA prohibits and how to apply it practically, our student data privacy guide covers every category in detail.

Part 2: Decide Which Tasks AI Drafts and Which Tasks You Own Entirely

Not every teaching task benefits from AI assistance, and not every task where AI could help should be delegated. Your framework needs a clear line between tasks where AI drafts the first version (and you review and edit) versus tasks where you produce the work directly.

A practical starting point:

AI drafts first, you review and finalize:

• Generic email templates and parent communication frameworks
• Rubric structures and assessment question banks
• Lesson plan outlines and unit frameworks
• Progress report comment drafts based on anonymized student profiles
• Administrative documentation that doesn’t involve student-specific data

You produce directly, AI doesn’t draft:

• Final evaluative judgments on student work
• Any communication where the specific relationship context matters (a difficult parent conversation you know the history of)
• Disciplinary records or behavioral documentation
• Anything that will be presented as your original professional assessment without disclosure

The key distinction: AI as a drafting assistant for structure and language is appropriate. AI as the source of professional judgment about a specific student is not.

Part 3: Establish Your Review Standard

Whatever AI drafts, you review. But “review” means different things depending on the task. A rubric draft that AI generates needs to be checked against your actual learning objectives — you know what your students need, the AI doesn’t. A parent email needs to be checked for tone, accuracy, and any context-specific sensitivity the AI couldn’t have known.

Write down one sentence for each task type you use AI for: “Before I use this AI output, I will verify _.”

For lesson plans: “I will verify that all activities are appropriate for my students’ actual readiness level and that the standard cited is correct.”

For feedback comments: “I will verify that the feedback reflects what I actually observed in this student’s work, not a generic description.”

For parent emails: “I will verify that the tone is appropriate for this specific relationship and that no details have been hallucinated or fabricated.”

This step adds two to five minutes per task. It’s the difference between using AI responsibly and outsourcing your professional judgment.

Part 4: Decide What You’ll Disclose and to Whom

This is the question most teacher AI frameworks skip — and it’s increasingly important. As AI use becomes visible (through AI detection tools, through parental awareness, through district audits), teachers who have thought through their disclosure position in advance are better prepared than those who haven’t.

A reasonable personal position for 2026:

With students: be honest that you use AI tools to help prepare materials, the same way you might use a template, a generator, or a textbook. You don’t owe students a detailed accounting of every tool you use, but if asked directly, honesty is both ethically appropriate and models the transparency you’re asking of them.

With parents: same principle. If a parent asks whether AI was involved in a progress report comment or a rubric, an honest answer is professionally appropriate. The position is that AI drafted a structure that you reviewed, edited, and finalized — which is accurate.

With administrators: if your district conducts an AI audit or asks teachers to disclose their AI use, having a written personal policy to share demonstrates intentionality, not carelessness. A teacher who can say “here’s my personal framework for AI use” is in a very different position than one who can’t explain their practices.

When Your School Finally Does Issue Guidance

RAND recommends that trusted sources — states, districts — provide guidance on what effective AI policies look like, updated regularly as the technology changes. When that guidance eventually arrives at your school, your personal framework becomes a foundation for engaging with it rather than a replacement for it.

Read new institutional guidance critically: does it distinguish between AI for administrative tasks versus AI for evaluative judgments? Does it address data privacy specifically, or just academic integrity? Does it give teachers clear safe-harbor rules for approved tools?

If the guidance is unclear, you now have a personal framework to compare it against — which puts you in a position to ask specific, informed questions rather than simply receiving policy passively.

The Practical Starting Point: One Week, One Decision

If building a full four-part framework feels like too much to do right now, start smaller. This week, make one decision and write it down:

“I will / will not use AI to draft [specific task]. When I do, I will always [specific review step] before using the output.”

That single decision, written down and followed consistently, is more valuable than an elaborate framework you never apply. Build from there.

For specific AI workflows that work within these principles — exact prompts you can use immediately — the copy-paste AI prompt guide covers the four highest-return tasks with templates that keep student data out of the equation.

If you’re finding that the ambiguity around AI use at your school is adding to your overall sense of exhaustion, that’s not just a technology problem — it’s a workload problem. Our teacher burnout guide covers the broader picture of what’s driving teacher depletion in 2026 and what actually helps.

Frequently Asked Questions

Is it my responsibility to create my own AI policy if my school hasn’t?

Not formally — policy is an institutional responsibility. But in the absence of institutional guidance, having your own personal framework protects you professionally and reduces the decision fatigue of navigating each AI use case individually. It’s practical self-protection, not extra work.

What happens if I use AI in a way that later turns out to violate a district policy that didn’t exist when I used it?

This is a genuine risk in a rapidly evolving space. The most protective position is one where your use is documented, defensible, and consistent with the four principles above — especially around student data. Acting in good faith with a coherent personal framework is a much stronger professional position than having no framework at all.

My district gave us a one-page AI policy that doesn’t actually answer any specific questions. What do I do?

A vague policy is better than no policy for institutional purposes, but it doesn’t help you make day-to-day decisions. Use the four-part framework above to fill in the gaps. If the policy doesn’t address data privacy specifically, apply the strictest standard: nothing that identifies a student goes into any consumer tool.

Are there AI tools that are safe to use without a district agreement?

Some tools are designed specifically for educational use and include built-in data protections — Google Workspace for Education (when your district uses it), ChatGPT’s Edu tier (with a district agreement), and purpose-built education platforms like MagicSchool AI that include FERPA compliance by design. Consumer-grade free tiers of general AI tools are where the data risk is highest.

---

This article is for informational purposes only and does not constitute legal or policy advice. Data privacy laws and district policies vary by jurisdiction and institution. Consult your district’s data privacy officer or legal counsel for guidance specific to your situation.

Sources

• Gallup / Walton Family Foundation — Teaching for Tomorrow: Closing the Expectations Gap (May 2026)
• RAND Corporation — AI Use in Schools Is Quickly Increasing but Guidance Lags Behind (September 2025)
• Education Week — Teachers Say Lack of AI Guidance Is a Major Problem (May 2026)
• K-12 Dive — Teachers Lack Formal AI Guidance for Learning and Instruction (May 2026)