Building an AI Steering Group: Roles, Responsibilities, and a Template Terms of Reference

Every school needs someone responsible for AI. Not in the vague, everyone-and-no-one sense that currently dominates, but in the structured, documented, accountable sense that applies to safeguarding, health and safety, and financial oversight.

The mechanism I recommend is an AI steering group: a small, cross-functional team with a clear remit, regular meeting cadence, and defined accountability to senior leadership and the governing body. This is not bureaucracy for its own sake. It is the governance infrastructure that turns an AI policy from a document into a living practice.

I have helped schools across the Middle East and beyond set up these groups, and in this article I am going to walk you through exactly how to do it, including a template terms of reference you can adapt for your own context.

Why a Steering Group, Not an Individual

Some schools appoint a single person as their "AI lead." This is better than nothing, but it has a fundamental problem: AI touches too many areas for one person to oversee effectively. It affects curriculum, assessment, data protection, safeguarding, IT infrastructure, professional development, and parent communication. No single individual has expertise across all of these areas, and expecting one person to manage them all is a recipe for either burnout or blind spots.

A steering group brings together representatives from each relevant area, ensuring that decisions about AI are informed by multiple perspectives. The data protection officer can flag compliance concerns; the head of teaching and learning can speak to pedagogical implications; the IT lead can address technical feasibility; a classroom teacher can ground discussions in practical reality.

This cross-functional approach also distributes ownership. When AI governance is one person's job, it is easy for the rest of the school to disengage: "that is Sarah's thing." When it is a group responsibility, with representatives from across the school, engagement is broader and decisions have more institutional buy-in.

Composition

Keep the group small: five to seven members is ideal. Large committees achieve nothing. Here is the composition I recommend:

Chair: A member of the senior leadership team. This signals institutional seriousness and ensures the group has the authority to make decisions and escalate issues. Ideally, this is the deputy head or the member of SLT responsible for academic strategy; not the IT director, because AI is an educational issue, not a technology issue.

Data protection officer or compliance lead. Every AI decision has data protection implications. Having this expertise at the table prevents the group from approving tools or practices that create compliance risk.

Head of teaching and learning (or equivalent). This person ensures that AI governance connects to pedagogical practice. They bring the curriculum perspective and can assess whether AI initiatives actually serve educational outcomes.

IT or digital learning lead. The technical voice. They understand infrastructure constraints, security requirements, and integration challenges. They also tend to have the earliest visibility of new tools being adopted across the school.

A classroom teacher. This is the most important voice in the room and the one most often missing. A practising teacher brings the reality check: what is actually happening in classrooms, what teachers actually need, what students are actually doing. Rotate this position annually to ensure diverse perspectives.

Optional: A parent representative. If your school culture supports it, including a parent representative ensures that community perspectives are heard. This is particularly valuable in international schools where parental expectations around technology, privacy, and education vary significantly.

Optional: A student representative (for secondary and sixth form). Students are the primary users of AI tools and often know more about what is happening in practice than staff do. Their voice adds a dimension that no adult in the room can replicate.

Remit

The steering group's remit should be clearly defined and documented. I recommend four core responsibilities:

1. AI tool approval. No new AI tool should be deployed with student data without passing through the steering group's review process. This does not mean the group conducts the technical evaluation; that is the IT lead's job. It means the group reviews the evaluation, considers the data protection implications, assesses the pedagogical value, and makes a collective decision.

2. Policy maintenance. The steering group owns the school's AI policy. It reviews the policy at least termly, updates it when circumstances change, and ensures it remains aligned with the school's AI strategy and the evolving regulatory environment.

3. CPD oversight. The group monitors the school's investment in teacher AI competency. It reviews CPD provision, examines competency data, and makes recommendations to SLT on where additional investment is needed.

4. Incident response. When an AI-related issue arises, an assessment integrity concern, a data protection incident, a parental complaint, the steering group convenes to assess the situation, determine the response, and identify any policy or practice changes needed to prevent recurrence.

Meeting Cadence

I recommend meeting half-termly during the academic year: roughly six meetings per year. This is frequent enough to maintain momentum and respond to developments, but not so frequent that it becomes burdensome.

Each meeting should follow a standard agenda: review of actions from the previous meeting, new AI tool submissions for approval, policy review items, CPD update, any incidents or concerns, and forward planning.

Minutes should be recorded and shared with SLT. A summary report should go to the governing body at least once per term, as part of the broader governance reporting cycle.

Template Terms of Reference

Here is a template you can adapt for your school:

Name: [School Name] AI Steering Group

Purpose: To provide cross-functional governance and oversight of AI use across the school, ensuring that AI tools and practices are safe, compliant, educationally valuable, and aligned with the school's strategic plan.

Membership: [List positions, not names; positions are more durable]

Chair: [SLT member responsible for academic strategy]

Quorum: A minimum of [three/four] members must be present, including the Chair or their designated deputy.

Meeting frequency: Half-termly during the academic year (six meetings per year), with provision for extraordinary meetings as required.

Core responsibilities:

• Review and approve all AI tools that process student data before deployment
• Maintain and update the school's AI policy, reviewing it at least termly
• Monitor teacher AI competency development and advise SLT on CPD priorities
• Coordinate the school's response to AI-related incidents
• Commission and review the school's annual AI readiness audit
• Report to the governing body on AI governance at least once per term

Decision-making: Decisions are made by consensus. Where consensus cannot be reached, the Chair has the casting vote. Decisions that have significant financial, legal, or strategic implications are referred to SLT for final approval.

Reporting: Minutes are shared with SLT within one week of each meeting. A termly summary report is prepared for the governing body.

Review: These terms of reference are reviewed annually at the first meeting of the academic year.

Making It Work

A terms of reference is just a document. What makes a steering group effective is culture and commitment; here is how to build it.

The Chair needs to protect the group's time and authority. If SLT overrides the group's decisions without consultation, the group becomes performative. If meetings are routinely cancelled or poorly attended, governance erodes.

The group needs access to data. The AI Literacy Audit Tool provides a structured, repeatable assessment that the steering group can use as its evidence base; running the audit annually gives the group a clear picture of progress and remaining gaps. But even without the tool, the group should have access to the AI tool register, CPD participation and competency data, and incident reports.

And the group needs to communicate outward. Teachers should know the steering group exists, what it does, and how to submit tools for approval or raise concerns. Parents should know that AI governance is being actively managed. The governing body should receive regular, substantive reports; not one-line reassurances.

AI governance in schools is not a solved problem. It is an ongoing function that requires sustained attention, cross-functional expertise, and institutional commitment. An AI steering group, properly constituted and properly supported, is the most effective mechanism I have seen for delivering that.