On Friday 10 April, school leaders, accreditors, technology partners and educators came together at the United Nations International School in New York for The Educators AI Playbook LIVE — a one-day strategy intensive designed to move the AI conversation beyond noise, fear and theory.
The premise of the day was simple: AI is moving too quickly for schools to rely on traditional planning cycles. The question is no longer whether AI will affect teaching, learning, operations, safeguarding, privacy, assessment and school strategy. It already is. The real question is whether schools will respond deliberately, ethically and systemically — or continue to react tool by tool, incident by incident, policy by policy.
From the opening session with Mark Orchison and Dan Fitzpatrick, through keynotes from Dan and Dr Sabba Quidwai, panels with NAIS, ATLIS, WASC, NEASC, CAIS and NJAIS, practical strategy sessions, student perspectives, Microsoft and Google, and hands-on workshops, one theme ran through the day:
AI leadership is not about chasing every new tool. It is about building the judgement, structures and confidence to decide what should change, what should stay human, and what schools are ultimately trying to protect.
Dan Fitzpatrick opened the day by confronting one of the biggest barriers facing schools: AI decision paralysis.
Many schools are waiting for certainty. They are waiting for perfect policy, perfect evidence, perfect tools, or perfect regulatory clarity. But AI is not waiting. The shift from simple chatbots to agentic AI — systems capable of pursuing goals, using tools, completing workflows and acting with increasing autonomy — is already changing what students, staff and school systems can do.
Dan’s challenge was not reckless adoption. It was the opposite. Schools need to stop confusing caution with inaction. The risks are real, but refusing to engage does not make those risks disappear. It simply means students and staff are left to experiment without shared expectations, without guidance, and without the structures needed to make good decisions.
His “three-box” framing offered a helpful way for school leaders to think strategically:
Box one is the present: how AI can improve current practice.
Box two is the past: what schools may need to stop doing because it no longer serves students in an AI-rich world.
Box three is the future: where schools can safely test new approaches to curriculum, assessment, workload, inclusion and learning design.
For many delegates, this was a useful shift. AI strategy cannot just be a list of approved tools. It has to include the more difficult leadership question: what assumptions about school are we willing to re-examine?
Dr Sabba Quidwai built on this by moving the conversation from fear to agency.
Her keynote challenged one of the most common contradictions in education. Schools are understandably worried about students using AI to cheat. At the same time, employers are increasingly clear that AI literacy, adaptability and the ability to work effectively with intelligent tools will be essential future skills.
This creates a mismatch. If schools define AI primarily as a threat to academic integrity, they risk treating the very capabilities students need as something to be avoided rather than developed. The task is not to ignore cheating. It is to design learning environments where AI use becomes visible, intentional and connected to deeper thinking.
Sabba’s “Move 78” metaphor captured this beautifully. In the famous AlphaGo match, the AI shocked the world with a move no human expected. But Lee Sedol’s response — “Move 78” — was a profoundly human act of creativity, resilience and strategic imagination. The lesson for schools is not that humans must compete with AI on its terms. It is that we must identify the uniquely human moves that matter most.
Those moves are rooted in agency, ambition and audacity.
Agency means helping students and staff understand how systems influence them, and how they can influence those systems in return.
Ambition means using AI to rethink problems, not simply to make existing tasks faster.
Audacity means being willing to redesign practice, scale pockets of excellence, and stop pretending that old models of assessment and workload will remain untouched.
One of the most powerful practical ideas from Sabba’s session was the importance of “prompting the human before the machine”. Before asking AI for an answer, schools should first ask: What is the situation? What problem are we trying to solve? What outcome would matter? What would make this more human, more effective, or more equitable?
That is the difference between AI as a shortcut and AI as a thinking partner.
The panel hosted by Debra Wilson, President of NAIS, helped ground these ideas in the reality of school leadership.
A recurring theme was that AI is forcing schools to re-examine how they define learning, effort and evidence. If an assignment can be completed by an AI system with little student engagement, the problem may not only be the tool. It may be the design of the task.
This does not mean abandoning rigour. It means asking better questions about where the learning happens. Is it in the final essay, or in the argumentation process? Is it in the answer, or in the questioning, critique, iteration and defence of thinking? Is assessment measuring completion, or is it measuring understanding?
The panel also surfaced a more empathetic reading of student behaviour. When students use AI badly, it is not always because they are trying to avoid learning. Sometimes they are overwhelmed, anxious, tired, under-supported, or unclear about what good use looks like. Schools need boundaries, but they also need instruction.
AI literacy is therefore not just technical. It is ethical, cognitive and relational. Students need to know how to question outputs, protect privacy, recognise bias, cite or disclose use, and decide when not to use AI at all.
The human must remain the architect of thought.
One of the most significant sessions of the day focused on the future of AI and accreditation, with leaders from ATLIS, WASC, NEASC, CAIS and NJAIS.
This conversation mattered because it moved AI out of the innovation silo and into the wider language of school quality, accountability and institutional maturity.
If AI is becoming part of teaching, learning, operations, student support and decision-making, then it will inevitably become part of how schools evidence responsible leadership. Accreditors are not simply interested in whether a school has an AI policy. They are interested in how strategy, governance, implementation, professional learning, risk management and impact fit together.
This is where many schools now need to mature their approach. A policy may be necessary, but it is not enough. Schools need to show how AI decisions are made, who is involved, how risks are assessed, how vendors are reviewed, how staff are trained, how students are engaged, and how leadership teams know whether AI is improving or undermining the school’s mission.
The emerging standard is not “have you adopted AI?” It is “can you evidence that you are governing it well?”
The middle of the day was deliberately practical.
The Strategy Buffet sessions tackled some of the most pressing questions schools are dealing with now: vendor evaluation, teacher-first AI roadmaps, AI-supported literacy, student perspectives and the uncomfortable questions leaders often do not have space to ask.
The vendor conversation was especially important. Schools are no longer just buying software. They are adopting systems that may include AI features, automated decision-making, data processing, content generation, monitoring, recommendation engines, or opaque third-party integrations. The phrase “AI-powered” is not a control. It is a prompt for due diligence.
Schools need to know what a tool does, what data it uses, how outputs are generated, what risks it introduces, whether children are affected, and how the vendor can evidence safety, privacy, transparency and accountability.
The Turing Trials workshop then brought these issues to life through scenario-based decision-making. Rather than discussing AI risk in the abstract, participants had to respond to realistic school situations: complaints about facial recognition, AI-generated grading disputes, safeguarding concerns, unclear vendor claims and competing stakeholder pressures.
This format mattered because AI governance is not a document exercise. It is a judgement exercise. Schools need leaders who can sit together, weigh benefits and risks, understand their roles, and make decisions that are defensible, values-led and practical.
The afternoon session with Google for Education and Microsoft Elevate added another important dimension: schools should not be passive recipients of AI tools.
The conversation moved beyond product demonstration and into partnership. If AI tools are going to shape the future of learning, educators need a voice in how they are designed, tested, implemented and improved, with the conversation acknowledging that this is not always easy. Large technology companies move at speed and scale. Schools operate in human, pastoral and community contexts where trust matters deeply. It was agreed by all that the sector needs more honest conversations from front line teaching on how AI products need to evolve to support teaching and learning. A great takeaway was “AI in education cannot be something done to schools. It has to be shaped with them.”
The day closed with a clear sense that schools do not need more hype. They need structure, confidence and action.
Five takeaways stood out.
First, AI strategy must be mission-led. Schools should not start with tools. They should start with the kind of learning, community and future they are trying to build.
Second, governance matters. AI decisions need clear ownership, cross-functional input, risk assessment, vendor due diligence and ongoing review.
Third, assessment must evolve. If AI can complete the task, schools need to rethink where the learning is evidenced.
Fourth, students need agency, not avoidance. AI literacy should help students make informed, ethical and purposeful choices.
Fifth, the human work of schools must remain central. AI should be used to strengthen relationships, judgement, creativity, inclusion and care — not to bypass them.
The Educators AI Playbook LIVE was not a day about predicting the future. It was a day about helping schools lead it.
The message was clear: stop waiting for perfect certainty. Build the structures. Ask the harder questions. Involve the right people. Test carefully. Learn quickly. Keep humans at the centre.
That is how schools move from AI paralysis to practical leadership.
We will be hosting future Educators’ AI Playbook events across the globe in 2026, Stay tuned for the announcements.