Most people who write about artificial intelligence (AI) seem to fall into one of three camps. AI is amazing. AI is terrifying. Or perhaps most commonly: the government needs to tell us what to do. None of these positions is particularly helpful. Artificial intelligence is already embedded in our lives. Schools and trust colleagues are using it to draft communications, analyse information and reduce administrative workload. Pupils are encountering it within education and beyond it. Suppliers are increasingly embedding it into products that we already use, often without organisations actively seeking it out. The question is no longer whether AI will affect our schools. It already does. The real question is how we respond. Certainty rarely arrives For some, the answer is to wait. Wait for regulation. Wait for guidance. Wait until the technology stabilises. Wait until somebody else has worked out the risks. But governance exists precisely because certainty rarely arrives. Trust boards routinely make decisions in conditions of uncertainty. They oversee safeguarding without being safeguarding specialists. They govern cyber security without being cyber security experts. They make decisions about finance, estates, procurement, workforce strategy, and organisational risk without needing to be professional accountants, surveyors, lawyers or human resources practitioners. AI is no different. That is because AI is not primarily a technology issue. It is a governance issue. This observation sits at the heart of Governing AI in school trusts, a new publication by the Confederation of School Trusts (CST) and TransforMATive. The challenge for boards is not deciding whether AI is good or bad. The challenge is ensuring that its use is safe, ethical, accountable and aligned with the trust’s educational purpose. That is fundamentally a governance question. Retaining human judgement Discussion focuses too much on the technology itself. Which tools should schools use? Which platforms are safest? Which safeguards are needed? These are important questions, but they are secondary. The more important question is this: How do we maintain human judgment when AI becomes ubiquitous? Every major technological innovation changes not only what we do, but how we think. AI can draft reports, summarise information, identify patterns, suggest actions, and generate recommendations at a scale and speed that would have been unimaginable a few years ago. These capabilities bring significant opportunities. CST’s national school trust report found that trusts are already using AI to reduce administrative burden, support operational efficiency, improve access to information and free up valuable staff time. The Department for Education has highlighted the potential for AI to reduce workload and improve efficiency too. But good governance requires us to ask a second question: What happens when convenience starts to replace judgment? Difficult questions The philosopher C. Thi Nguyen argues that tools and systems are never entirely neutral. They shape what we pay attention to, how we make decisions, and ultimately what we value. Our guidance draws on this insight because one of the greatest governance risks associated with AI is not necessarily that it will make decisions for us. It is that we gradually stop exercising our own judgment. That risk is particularly important for trustees. Good governance depends on asking difficult questions, weighing competing priorities, understanding context, exercising discretion and making decisions in the best interests of children and young people. AI can provide information. But it cannot reconcile competing values. It cannot decide what level of risk is acceptable. It cannot exercise accountability. Nor can it tell a board what kind of organisation it wants to be. Those responsibilities remain firmly with people. This is why trustees do not need to become AI experts. They need to be good trustees. Resilient principles The principles that underpin effective trust governance remain remarkably resilient, even in the face of rapid technological change. Educational purpose still comes first. Public money still needs to be spent wisely. Safeguarding still matters. Accountability still matters. Stakeholder trust still matters. Boards still need evidence, assurance, transparency and effective risk management. The governance challenge is therefore not to create an entirely new model for governing AI. It is to apply enduring governance principles to a new and fast-moving context. This means asking sensible questions about how and where AI is used, who is affected, what training do we need and where should human judgment remain non-negotiable? These are governance conversations, not technical conversations. That distinction matters because governance is one of the few disciplines specifically designed to help organisations navigate uncertainty. Boards rarely have the luxury of complete information before acting. They need enough information to exercise informed judgment, supported by clear values, effective assurance and proportionate oversight. There will undoubtedly be new risks, opportunities and questions as AI continues to develop. But the core responsibility of governance will remain the same. Not to eliminate uncertainty. Not to wait for perfect certainty. But to steward organisations wisely through it. AI may be new. The principles of good governance are not. And that is precisely why school trusts are better prepared for this challenge than many people realise.