AI governance in MATs is no longer a future issue. In many trusts, AI tools are already being used in planning, assessment, administration, and wider decision-making.
The challenge for leaders is not simply whether AI is being used, but whether boards have enough visibility, evidence, and accountability to oversee that use with confidence.
That is why the launch of the AI Governance in MATs survey matters. It has been created to help build the first sector-wide benchmark of how multi-academy trusts are governing AI, managing risk around children’s data, and giving boards meaningful assurance. At a time when expectations around oversight are rising, MAT leaders need more than general guidance. They need a clearer sense of what good practice looks like across the sector.
Recent education and data protection guidance has reinforced the same message: where AI use affects children’s data, organisations need human oversight, clear accountability, and evidence-based compliance. Yet one pattern continues to appear in conversations with MAT leaders. AI is often treated as an operational or technical issue, while assurance remains a governance responsibility.
That gap creates risk.
The survey behind the 2026 Benchmark Report is designed to make that gap easier to see. By gathering anonymised responses from MAT leaders and governors, it will build a more realistic picture of how trusts are approaching AI governance today and where boards may still need stronger oversight.
MAT leaders and governors can use the survey to benchmark their current AI governance arrangements, identify gaps in board assurance, and contribute to the first sector-wide picture of how trusts are managing AI-related risk.
For many trusts, the starting point is simple: understanding where AI is already in use and what that means for risk. Generative AI, predictive tools, and automated workflows may already be influencing classroom practice, internal administration, communication, or analysis. Without a structured view of that activity, it becomes difficult for leaders to judge whether existing controls are proportionate.
That is especially important when children’s data is involved. A low-risk use case may involve staff using AI to support planning without entering personal data. A much higher-risk use case might involve pupil information, behavioural profiling, automated recommendations, or third-party systems processing sensitive data on behalf of the trust. These are not abstract concerns. They directly affect safeguarding, accountability, and the ability to demonstrate lawful, well-governed practice.
A practical approach is to map AI use against a small number of clear questions: what data is involved, what decisions or outputs the tool influences, how much automation is involved, and what level of supplier dependency exists. This gives leaders a stronger basis for prioritisation and gives boards a clearer view of where scrutiny should sit.
This is also exactly the kind of issue the sector benchmark is designed to surface. The survey will help show whether trusts are moving from isolated experimentation to more structured oversight, or whether risk mapping is still inconsistent across the sector. By showing whether trusts are moving from isolated experimentation to structured oversight, the survey will help leaders see whether AI risk mapping is becoming embedded practice or remains inconsistent across the sector.
For boards, good AI governance is not mainly about policy wording. It is about evidence. Trustees need to be able to see where AI is being used, what controls are in place, what risks have been identified, and how those risks are being monitored over time. Without that visibility, assurance remains too dependent on informal updates or broad reassurance.
In practice, this may include a trust-wide register of AI or high-risk digital systems, linked where appropriate to existing Records of Processing Activity, Data Protection Impact Assessments (DPIAs), supplier due diligence, and internal approval routes. It may also include clearer reporting on training, use cases, incident response planning, and follow-up actions where risks have been identified.
Boards do not need operational detail on every tool. They do, however, need enough structured information to answer straightforward but important questions:
Can we see where AI is being used across the trust?
Do we have evidence of review, oversight, and accountability?
Are we confident that children’s data is being handled in line with our obligations and values?
The 2026 Benchmark Report is intended to provide a wider reference point for those questions. It will help trusts understand how consistently this kind of evidence is being maintained across the sector and where common gaps in board assurance are starting to emerge.
One of the biggest challenges for MAT leaders is not simply knowing what to do, but knowing whether their current approach is proportionate, mature, and keeping pace with change. Without a wider benchmark, it is difficult to judge whether AI governance arrangements are robust, still developing, or too informal for the level of risk involved.
That is why a sector-wide benchmark matters. It gives executive leaders, governance leads, and boards a clearer reference point for what effective oversight looks like across the MAT sector. It also helps identify where common pressure points are emerging, whether around visibility, approval processes, supplier due diligence, staff understanding, or the governance of children’s data.
Importantly, the value of the benchmark is practical, not theoretical. It can help trusts see whether they are broadly in line with peers, where they may need to strengthen accountability, and which areas should be prioritised over the next 12 months.
In that sense, the survey is not just collecting views. It is helping to create a stronger basis for board-level decision-making.
Once AI use and risk are visible, the next step is practical oversight. In many MATs, this will mean reviewing whether current data protection, safeguarding, digital strategy, and approval processes properly reflect the pace and nature of AI adoption. Controls need to be clear enough to guide day-to-day decisions, not simply sit in a policy library.
That may include setting approval expectations for new AI tools, clarifying when a DPIA is needed, defining who reviews higher-risk use cases, and making sure supplier arrangements are properly understood. It may also mean giving leaders a better route to monitor changes over time, especially where suppliers introduce new AI functionality after procurement.
Culture matters as much as policy. Staff need confidence to raise questions, flag unexpected behaviour, and ask for support without feeling they are slowing innovation down. Leaders also need enough shared language to explain AI governance clearly to boards, senior teams, and families in Plain English.
The survey explores these practical realities, not just headline intentions. That matters because the strongest governance approaches are usually the ones that connect policy, oversight, and day-to-day practice rather than treating them as separate issues
The AI Governance Survey in MATs takes around 10 minutes to complete and is designed for MAT leaders and governors who want a clearer picture of their current position. It focuses on five key themes: AI Governance, Board Oversight, Risk Management, Children’s Data, and Leadership Assurance.
Together, these themes reflect the areas where scrutiny is most likely to increase and where evidence-based compliance matters most. They also reflect the questions many boards are already beginning to ask:
Where is AI being used
What risks are we carrying
What evidence supports our oversight
Where should we focus next?
Because the survey is anonymous, it gives trusts space to answer honestly about areas that are still evolving. That honesty is essential if the sector is to build a benchmark that reflects real conditions rather than idealised practice. The stronger and more representative the response base, the more useful the benchmark will be for everyone involved.
When the Benchmark Report is published, trusts will be able to compare their current approach with anonymised sector patterns. That insight can help leaders identify strengths, spot gaps, and plan the next stage of their AI governance work with greater confidence and accountability.
As AI use continues to develop across education, MAT leaders need more than isolated examples and high-level commentary. They need a stronger evidence base for what board-ready oversight looks like in practice. This survey is an opportunity to help build that picture for the sector.
If your trust is working to strengthen visibility, accountability, and assurance around AI, taking part will do more than contribute to a report. It will help create the first sector-wide benchmark for AI governance in MATs and provide a clearer reference point for future decisions around children’s data, risk, and governance.
To learn more about the wider context of AI and data protection in education, explore existing guidance from GDPRiS.