← All policies

ExelAcademy and the AI Age

Our ethos, vision and commitment to Human Authority: how we prepare every learner to lead, not follow, in an AI world.

Version 1.0 · August 2026

Contents
  1. 1. Our Conviction
  2. 2. Why This, Why Now
  3. 3. Our Age Commitment
  4. 4. The Human Authority Framework™ in Practice
  5. 5. Teaching Leadership Over AI: A New Standard for a New World
  6. 6. Safeguarding Woven In, Not Bolted On
  7. 7. Our Promise to Families
  8. 8. Alignment to Global Standards

1. Our Conviction

Artificial intelligence is no longer a tool learners will meet one day. It is already inside their world: in the way information is found, the way questions are answered, and, increasingly, in the way work itself gets done. This is not a passing feature of modern life. It is a permanent shift, and it is happening during the exact years our learners are forming their habits of mind.

ExelAcademy's founding conviction is that this moment calls for a different kind of education. Not one that treats AI as a threat to be filtered out, nor one that hands learners over to it uncritically, but one that deliberately builds the human capability to direct it, question it, and remain accountable for what it produces. We call this Human Authority.

“Technology exists to extend human potential. It must never replace human purpose.”

“Human agency describes our capacity to choose. Human Authority describes our responsibility for those choices.”

2. Why This, Why Now

For most of the last two decades, AI in education meant a tool a learner used: a search engine, a spellchecker, a chatbot that answered a question. That era is ending. AI is rapidly becoming agentic: systems that do not just answer questions but carry out multi-step tasks, make recommendations, and act with a degree of independence, in the same way a member of staff might once have been asked to.

This changes what young people need to learn. The skills once reserved for managing people, giving clear instructions, checking work, questioning assumptions, and taking ultimate responsibility for the outcome, are becoming the skills every person will need in order to direct AI agents well. A learner who has never practised this will not simply be behind. They will be in the habit of accepting whatever the machine gives them, at exactly the point in history when that habit is most dangerous.

UNESCO's AI Competency Framework for Students (2024) makes the same case at an international level. It places human agency at the centre of AI education, stating that students must understand their responsibility to protect that agency throughout the design, provision, and use of AI, and that human choice should not be ceded to AI when making high-stakes decisions. ExelAcademy was built around this principle before we had the citation for it. Now we have both.

3. Our Age Commitment

ExelAcademy draws a clear and deliberate line between two different things: learning about AI, and learning independently through AI.

3.1 Learning About AI: From the Start

AI literacy begins early. From Foundation onwards, learners are introduced to what AI is, how it works, and how to think about it, through age-appropriate courses and, in the earliest years, unplugged and low-tech activities rather than direct interaction with a live system. Technology use is also woven into ExelQuests™ from an early age: learners are encouraged to use age-appropriate tools within their real-world projects, always human-facilitated, always with a Mentor or facilitator present, and never as a substitute for that human relationship.

3.2 Learning Independently Through an ExelTutor™: Not Before Thirteen

What does not happen before age thirteen is independent, unaccompanied learning through an ExelTutor™: a learner alone with an AI system, receiving personalised instruction with no human mediating that exchange. This is the threshold UNESCO recommends in its 2023 Guidance for Generative AI in Education and Research, which proposes a minimum age of thirteen for independent conversations with generative AI platforms. Thirteen is also the youngest age at which data protection law lets a child consent to online services for themselves: the EU's GDPR sets that age at sixteen but allows each country to lower it to no younger than thirteen, and the UK has set it at thirteen. ExelAcademy's Phase architecture was built to this standard from the outset: full ExelTutor™-supported study begins at age thirteen, within Orientation, alongside ExelQuests™, in line with our 20/80 model.

This is not a technology limitation. It is a considered position: a young person should be fluent in what AI is, and should have practised using it under guidance, well before they are ever given the independence to learn from it alone.

4. The Human Authority Framework™ in Practice

Every part of ExelAcademy's approach to AI sits inside the Human Authority Framework™. AI enhances the acquisition of knowledge; it is human expertise that transforms that knowledge into wisdom, through experience, reflection, and purposeful action. This is not a slogan we apply after the fact. It shapes what we build.

“What kind of education becomes possible when Artificial Intelligence assumes responsibility for much of the routine acquisition, reinforcement and personalisation of knowledge?”

“AI enhances the acquisition of knowledge; human expertise transforms knowledge into wisdom through experience, reflection and purposeful action.”

Within our 20/80 model, the 20 per cent of learning delivered through an ExelTutor™ is not a passive stream of content. It is the first place learners practise Human Authority in action: reviewing what the ExelTutor™ offers, checking it, and deciding what to do with it. The 80 per cent, delivered through ExelQuests™, is where that same discipline is rehearsed against real stakes, real projects, and real outcomes.

5. Teaching Leadership Over AI: A New Standard for a New World

What follows is not a single lesson or a nice-to-have enrichment activity. It is ExelAcademy's position that leadership over AI has become a core standard of a modern education, in the same category as literacy and numeracy, and it sets out why we believe that, what we are seeing change in the world our learners are graduating into, and how we are building the standard into the design of ExelAcademy itself.

5.1 Why This Is Now a Core Standard, Not an Elective Skill

For most of educational history, the question a school had to answer was what a young person should know. That question has not disappeared, but a second one now sits alongside it: what a young person should be able to direct, check, and take responsibility for, when a machine can produce a fluent, confident answer to almost anything in seconds. A learner who has only ever been taught to find information has not been taught for the world they are entering. A learner who has been taught to interrogate, verify, and direct that information has.

This is not a speculative concern. The World Economic Forum's Future of Jobs Report 2025 found that employers expect close to two in five core workplace skills to change by 2030, with AI and big data literacy rising faster than any other skill category, and leadership and social influence entering the top ten fastestgrowing skills for the same period. Those two trends are not separate. As routine cognitive work is increasingly delegated to AI, the skills that remain distinctly valuable are the ones involved in directing, judging, and taking accountability for what AI produces, precisely the skills of leadership, applied to a new kind of subordinate.

5.2 What We Are Anticipating

The shift from AI as a tool a person uses to AI as an agent that acts on a person's behalf is already well underway in the adult workplace our learners will enter. Industry analysis from Gartner projects that taskspecific AI agents will be embedded in around 40 per cent of enterprise applications by the end of 2026, up from under 5 per cent at the start of 2025. Separate research from PwC and other major analysts puts the proportion of companies already adopting AI agents in some form at close to four in five, with enterprise deployment continuing to accelerate through 2026 as organisations move agents from pilot projects into daily operational use.

We take from this a clear and, we believe, urgent conclusion. Today's learners will not simply use AI as a search tool the way earlier generations used the internet. Many of them will graduate into workplaces where directing one or more AI agents, briefing them, checking their output, catching their errors, and remaining accountable for the result, is a basic feature of almost any role, not a specialism reserved for technologists. A young person who has never practised this before entering the workforce will be starting from zero at the exact moment their employer expects fluency. A young person who has rehearsed it since Foundation will not.

5.3 Our Standard

ExelAcademy therefore treats leadership over AI as a standard, not an aspiration, and holds every ExelTutor™ interaction and every relevant ExelQuest™ to it.

1. Verify before you accept. No AI output, however confident or fluent, is treated as true until it has been checked. Learners are taught to look for what is missing, invented, or simply wrong, as a matter of instinct, not as an occasional exercise.

2. Direct, don't defer. Learners practise giving AI clear, specific instructions and reviewing what comes back critically, in the same way a good manager briefs and reviews the work of a team member, rather than accepting the first output as finished.

3. Own the outcome. Responsibility for the accuracy, quality, and ethics of anything produced with AI assistance sits with the human who directed it. An AI agent's mistake is never treated as an excuse. 4. Judgement stays human. Wherever a decision is genuinely high-stakes, whether academic, ethical, or personal, the final judgement rests with a person, never with the AI system.

This standard is deliberately modelled on the human-centred mindset at the core of UNESCO's AI Competency Framework for Students, which states that students must protect their own agency throughout the design, provision, and use of AI, and that human choice should never be ceded to AI on high-stakes decisions. We have simply made it operational, and built it into how our learners actually spend their time.

5.4 How We Teach It: The AI Detective Quest

One example of this standard in practice is the AI Detective Quest. Learners are placed in the role of a detective investigating a case built from AI-generated material, and nothing they are given is to be taken at face value. They treat every AI output the way a good detective treats a witness statement: a claim to be tested, not a fact to be absorbed. They cross-reference sources, look for what has been left out or quietly invented, and learn to recognise the confident, fluent tone an AI system uses even when it is wrong. A related strand places learners in direct leadership of an AI agent: giving it a brief, reviewing what it returns, catching errors before they matter, and deciding what is fit for use and what must be redone. As with any other quest, the learner, not the tool, is marked on the outcome. These are illustrative examples of the standard in Section 5.3, not the standard itself; further Quests built to the same standard will follow as our curriculum develops.

5.5 Essential Skills™ at the Centre

This standard sits inside our wider Essential Skills™ provision: the capabilities we consider critical to human relevance in the AI age, including judgement, communication, accountability, and the confidence to question a confident-sounding answer. These are not soft add-ons to an academic curriculum. They are, on the evidence set out above, becoming the load-bearing skills of the next working generation, and we intend for every ExelAcademy learner to leave us fluent in them.

6. Safeguarding Woven In, Not Bolted On

Teaching leadership over AI only works if learners are also protected from the ways AI can quietly undermine that leadership. ExelTutors™ are built so that they never model the kind of relationship a learner should be questioning: they do not offer friendship, do not encourage secrecy, and do not use emotionally bonding language. Human relationships, with Mentors, facilitators, and family, remain the primary relationships in a learner's life, and every ExelTutor™ is designed to reinforce that, not compete with it.

This reflects growing international concern, including UNICEF's updated 2025 guidance on AI and children, which specifically highlights the risks of unsafe AI companions and emotional over-reliance on AI systems by young people. ExelAcademy's full technical safeguarding requirements, covering privacy, content moderation, crisis escalation, and the guardrails imposed on any technology partner, are set out separately in our AI Tutor Safeguarding Standards and Development Guardrails Policy. This document sets the ethos; that document sets the enforceable detail.

7. Our Promise to Families

  • Your child will learn what AI is, and how to think about it, from their earliest years at ExelAcademy.
  • Your child will not learn independently through an AI tutor before the age of thirteen, in line with UNESCO's recommended threshold.
  • Every use of technology before that age happens through ExelQuests™, under human facilitation, never as a substitute for a Mentor or family relationship.
  • From thirteen onwards, your child will be taught, deliberately and repeatedly, to question, verify, and take responsibility for what AI produces, not simply to accept it.
  • No ExelTutor™ will ever be designed to be your child's friend, confidant, or emotional substitute for the people in their life.

8. Alignment to Global Standards

This ethos is ExelAcademy's own, but it is not formed in isolation. It draws on and aligns with the following international frameworks.

  • UNESCO Recommendation on the Ethics of Artificial Intelligence (2021). The foundational global framework on AI ethics, emphasising human oversight and human-centred values.
  • UNESCO Guidance for Generative AI in Education and Research (2023). The source of our age-thirteen threshold for independent use of generative AI tools in an educational setting.
  • UNESCO AI Competency Framework for Students (2024). The framework underpinning our approach to human agency, critical judgement, and leadership over AI as a taught competency, not an incidental skill.
  • UNICEF Policy Guidance on AI and Children, Version 3 (2025). Informs our approach to emotional-dependency safeguards and the risks of AI companions.
  • OECD AI Principles. Inform our wider commitment to transparency, accountability, and human oversight in any AI system used at ExelAcademy.
  • World Economic Forum, Future of Jobs Report 2025. The source of our labour-market analysis in Section 5, including projected change in core workplace skills and the rising importance of AI literacy and leadership by 2030.
  • Industry analysis on agentic AI adoption (Gartner, PwC, and related 2025–2026 enterprise research). The source of our analysis of how quickly AI agents are being adopted into everyday workplace roles, underpinning our case for teaching leadership over AI agents from an early age.