For school principals, heads of department and AI committees

Who is accountable for AI in your school?

School AI Governance Program. From an unmade decision to a policy your staff can apply.

Free policy templates have existed since 2023, and classroom use of AI has grown faster than school policy has followed it. We measure capability against international standards, then close the gap.

Take the four-minute diagnostic Book a scoping conversation

The program takes a school from no position on AI to a written, communicated, enforceable policy, in six stages and roughly twelve weeks, with a measured baseline of what its teachers can actually do. The American Institute of Business & Technology runs the process and the measurement. The school makes the decisions.

Three gaps

Visibility

A practice nobody has measured

You cannot govern a practice you have not measured.

Liability

A policy nobody has written

A written policy converts an ambiguous situation into an accountable one, which is why it does not get written.

Ownership

A committee without an owner

A committee is what you build after there is an owner, not instead of one.

The evidence

Spring 2025 surveys of US schools, from RAND:

54%

of students used AI for school.

53%

of English, math and science teachers used AI for school.

45% vs 34%

of principals reported an AI policy, against the share of teachers who reported one. The policy exists on paper before it exists in the classroom.

80%+

of students said no teacher had explicitly taught them how to use AI.

1 in 2

students worried about being falsely accused of cheating with AI.

18%

of principals reported any school or district guidance on AI a year earlier, in spring 2024 (RAND).

In Lebanon, the CRDP framework is published and requires teacher AI training before classroom deployment, but sets no school-level mandate and defines no owner, so the decision sits entirely with the school. Schools with European curriculum partners, examination boards or platform contracts also work under the EU AI Act: its AI literacy duty has applied since February 2, 2025, its transparency duties since August 2, 2026, and its high-risk obligations for education were deferred to December 2, 2027 by Regulation (EU) 2026/1744.

Four decisions only the school can make

The data red line

Which information may never enter a third-party AI system.

Typically decided by: The head, with the records or data lead. Signed off by the board or governing body.

The disclosure threshold

At what point students must declare AI use.

Typically decided by: Heads of department jointly, by grade band. Signed off by the head.

Assessment validity

Which assessments still measure something with AI access.

Typically decided by: Each head of department, for their own subject. Signed off by the head.

The standard of evidence

What is sufficient to accuse a student of AI misconduct.

Typically decided by: The head, with the pastoral or discipline lead. Signed off by the board or governing body.

Everything else can be copied.

How the program works

Six stages in roughly twelve weeks. The sequence is a design judgment, not a measured finding, and the timings are indicative.

Stage 0
Set the default position
One meeting. Leadership picks one of three postures: permitted with disclosure, restricted by approval, prohibited pending review.
Week 1
Stage 1
Name one accountable owner
One person with decision rights and a budget line. A committee is what you build after there is an owner, not instead of one.
Weeks 1 to 2
Stage 2
Establish two baselines
A tool inventory including AI already embedded in the LMS, assessment platform and student information system, and an educator competency baseline, measured with an instrument, not self-reported.
Weeks 2 to 4
Stage 3
Decide the four. Copy the rest.
Ninety percent of a policy comes verbatim from free templates. Four decisions cannot: the data red line, the disclosure threshold, assessment validity, the standard of evidence.
Weeks 4 to 6
Stage 4
Communicate, then train, then enforce
In that order.
Weeks 6 to 10
Stage 5
Review on triggers, not on dates
The policy is reopened by its owner when something changes: a new tool or a vendor switching AI on, an incident or a parent challenge, a curriculum, regulatory or examination board change.
Week 10 onward

What the school receives

Three components, scoped to the school. No packages and no published prices.

Decide

Facilitation of Stages 0 and 1, and a half-day four-decisions workshop with the leadership team and heads of department. It produces the school’s written positions on the four decisions and a decision-rights matrix: who decides, who is consulted, who signs off.

Measure

The educator competency baseline, structured against the UNESCO AI Competency Framework for Teachers and mapped to the Lebanese CRDP teacher framework. Re-measured after development, so improvement is evidenced rather than asserted.

Implement

The policy drafted from a free template with the school’s four decisions inserted, a communication sequence for families and students, staff training targeted by the baseline, and a trigger list with a named owner for review.

The baseline returns:

An individual profile for each educator.
A departmental view.
An institutional baseline figure a board or accreditation panel will accept.
A ranked gap list.
A development plan by cohort.

The purpose is not to classify teachers. It is to know what development the school needs before it buys any.

Departmental baseline viewIllustrative

UNESCO dimension	English	Math	Science	Arabic
Human-centred mindset
Ethics of AI
AI foundations and applications
AI pedagogy
AI for professional development

One segment is Acquire, two Deepen, three Create. Each cell is the department’s median level on that dimension.

Illustration. The dimensions and levels are the framework’s; the departments and values are invented and describe no real school.

What we will not do

We do not make the four decisions for the school.
We do not sell a template. The good ones are free, and we link to them below.
We do not use AI detectors as a standard of evidence.
We do not publish school-level data.

Free resources

None of this is behind a paywall. Adoption is not an access problem.

TeachAI AI Guidance for Schools Toolkit
Common Sense AI Toolkit for School Districts
EDSAFE SAFE Benchmarks Framework and the EDSAFE AI Policy Stack
Digital Promise acceptable-use language
UNESCO: the teacher and student AI competency frameworks, and the guidance for generative AI
EU AI Act: Article 4 and Article 50

Where the school sits

The diagnostic tells you where your school sits, not where we sit.