From Policy to Practice for Secondary Schools
Most school AI policies are directionally fine and operationally weak. The real work begins when policy language has to become teacher moves, student expectations, and family-facing explanations.
Many schools now have an AI policy, a guidance document, or a statement of principles.
That is progress, but it is only early progress.
Most policies fail at the same point: they say what a school believes, but they do not tell teachers what to do on Monday.
Secondary schools feel this gap especially sharply because their implementation problems are not generic. A principal, an AP teacher, a department chair, and a ninth grade humanities team all need different levels of translation.
What strong policy documents usually get right
Most emerging policies do a reasonable job on first principles:
- they acknowledge both opportunity and risk
- they name privacy and academic integrity concerns
- they avoid framing AI as purely good or purely bad
- they ask for human judgment
That is the easy part.
The hard part is operationalizing those values across a real school.
Where policy usually breaks down
In my experience, schools get stuck in one of four places.
1. The language stays abstract
Teachers get broad statements like "use AI ethically" or "maintain academic integrity," but no examples of what acceptable classroom use looks like in practice.
2. The guidance is too centralized
Leadership teams write a thoughtful document, but departments are left to interpret it alone. That creates uneven implementation and mixed student messaging.
3. Families are brought in too late
If parents hear about AI only after confusion or conflict, the school has already lost trust it could have built earlier with clear examples and rationale.
4. Assessment remains unchanged
Schools revise policy language while leaving the core assessment model intact. If the work still rewards polished final outputs more than evidence of thinking, the policy will not hold.
What better rollout looks like
A usable policy rollout has to create artifacts, not just beliefs.
For secondary schools, I think the minimum viable set includes:
- a teacher-facing one-page summary by scenario
- student-facing classroom norms
- family communication that explains the school's logic in plain language
- sample assignment language
- a short rubric or evidence guide for AI-supported work
These artifacts help different groups interpret the same policy coherently.
Start with the scenarios that create the most friction
Do not begin with edge cases. Begin with the recurring questions teachers are already asking:
- Can students use AI for brainstorming?
- What counts as original thinking in a drafted essay?
- How should AI use be disclosed?
- What evidence do we want in project-based work?
- What does acceptable use look like in AP or honors contexts?
If policy documents do not answer those scenarios, teachers will answer them individually.
That may be unavoidable in some cases, but it should not be the default operating model.
Build the teacher version first
One of the biggest mistakes I see is over-investing in polished leadership language before producing teacher-facing implementation support.
The teacher version should be concrete:
- sample statements you can paste into assignments
- examples of allowed and not-allowed use
- suggested checkpoints for documenting thinking
- department-specific prompts for calibration
Once teachers have that clarity, student and family communication becomes much easier.
Policy must connect to professional learning
A static document does not build capacity.
If a school wants coherent implementation, it needs some combination of:
- calibration sessions using real assignments
- examples from different subject areas
- shared discussion of evidence and assessment
- follow-up after initial rollout
Without that learning loop, the policy becomes another file in a shared drive that everyone references and no one fully uses.
The most important shift
The biggest shift is this: schools need to stop thinking of AI policy as a document and start treating it as an implementation system.
That system includes language, examples, training, assessment, and communication. If any one of those pieces is missing, the whole thing becomes harder to trust.
For grades 9-12, where rigor and consistency matter, that trust is everything.
Policy is necessary.
But practice is what students actually experience.
Signature framework
The Assignment Design System
See the full assignment design system that connects the ideas in this article to classroom practice.
Resource engine
Move from reading to action
Use the readiness quiz, lesson analyzer, policy builder, and assignment library to put the ideas into practice.

Chris Meehan
Academic Technology Director at Berkshire School, researching AI in grades 9-12 at Brown University. I publish practical frameworks, tools, and articles for secondary-school educators navigating AI.