As long as society continues to judge young people on the basis of these outputs, schools will rightly continue to prioritise them. And yet, if education is treated purely as an extrinsic good, valuable only as a means to the ends of academic qualifications, university places, and well-paid careers, we risk focusing on output rather than development and overlooking the intrinsic value in the educational process itself.

This reductive view creates particular challenges for schools in their response to Artificial Intelligence. When confronted with generative AI, the first instinct of many teachers and academics is reactive: this could be used for cheating; pupils won’t learn to think; let's ban it! AI is undeniably disruptive to established patterns of school life, a threat to familiar institutions like homework and coursework.

We cannot, however, afford to bury our heads in the sand. As former Prime Minister Rishi Sunak observed, “You are more likely to lose your job to someone who is using AI, than to AI itself." This view is reinforced by research from PwC, which indicates that workers with AI fluency already command a substantial wage premium. If schools are to prepare young people to thrive in the world, they have a duty to engage with it constructively and to develop genuine fluency in its use.

However, the emergence of what is increasingly referred to as a cognitive industrial revolution does not herald a retreat from cognitive excellence. Rather, it requires us to rethink what cognitive excellence entails. This is not unprecedented. Before the invention of the printing press, cognitive excellence was closely associated with the ability to memorise information. The mass production of texts shifted that emphasis. The internet revolution brought a similar recalibration. In each case, new technologies did not diminish the need for cognitive excellence but changed the nature of it. The emergence of AI is arguably no different.

Generative AI is a tool; one that I confess to using, from time to time, in the drafting of this blog. The crucial question is not if you use AI, but how. I employ it not as a shortcut, but as a means to improve the overall quality of the final product. It rarely saves me time; rather, it can sharpen the quality of the writing (in particular the spelling and grammar) and, more occasionally, help to test the clarity and coherence of an argument. The ideas, however, are my own; accumulated, synthesised, and developed from things read, podcasts listened to, programmes watched and conversations with interesting people. Initially scribbled down in notebooks or captured in the Notes app on my phone, before being fashioned into a ‘rough’ draft, which is then developed through editing and judicious use of AI. I do not delegate thinking, and suggestions from Copilot are frequently rejected. The finished product is my work; my ideas, expressed in my words.

Would I be content for our students to use AI in the same way? To improve the quality of, for example, a history essay?

The answer is far from straightforward. Context is vital, integrity even more so. Clearly, it would not be acceptable for pupils to use AI to generate a piece of coursework. Coursework is intended as a test of a candidate’s ability to produce an original response to a stimulus question, not a test of their ability to use AI. To present AI-generated work as one’s own would be dishonest. Honesty is more important than situation. Pupils should not use AI in any situation where it has been prohibited and where it is permitted, they must be transparent about it.

However, the ability to use AI well through thoughtful, well-constructed prompts, critical evaluation, and careful editing, is undoubtedly a skill young people must develop. Unfortunately, too often students see AI not as a tool to enhance their own work, but as a ‘cheat code’: a means of completing tasks with minimal effort.

But even when pupils use AI the right way, as a supplement rather than a substitute for thinking, there remains one crucial distinction between this blog post and a history essay. A distinction not of audience, subject matter, or assessment, but of purpose. In one case, the output is the objective; in the other, it is secondary to the process and that difference must shape how, and indeed whether, AI is part of that process.

Years of learning, research and writing - at school, university and throughout my career – largely undertaken without the aid of Gemini or Google (yes – I am old enough to have graduated university using a library rather than the internet as my primary research tool) have enabled me to develop the knowledge, reasoning and judgement necessary to deploy AI to enhance, not replace my work. Arguably the greatest threat AI poses to education is not that it allows people to cheat on homework and coursework by passing off the work of ChatGPT as their own, but that it allows them to bypass the very learning that would equip them to use it effectively.

True learning resides in the process, not merely the production of outputs. The point is to have written a history essay, not simply that one exists. As Theo Baker, a Stanford Computer Science graduate, put it in The New York Times: ‘Sure a robot can lift 600 pounds much more easily than I can – but that doesn’t help me much when I am trying to workout.’

As it becomes increasingly difficult to determine whether pupils have used AI, we must ensure they understand this important distinction. If they are to become the thoughtful, capable users of AI the cognitive industrial revolution requires and who will enjoy a competitive advantage in the labour market, they must recognise that education is not an industrial process designed to produce pre-determined outputs as efficiently as possible, but a journey on which the challenges they face and the hard work they commit will mould them.

Only when pupils recognise this and schools can trust them to use AI to enhance their educational outcomes rather than as a shortcut, are we truly free to implement the balanced response that this moment requires. A curriculum that exposes pupils to these powerful tools and provides opportunities to use them, while still requiring them to walk the miles through which they learn to use them well. A curriculum that develops not only prompt literacy, but also creative and critical thinking, relational intelligence and character. A curriculum that commits to both these strands intentionally and in equal measure.

To reach this point, we must change our messaging. We cannot reasonably persist with the language of outputs and extrinsic value and continue to be shocked when young people seek shortcuts to achieve them. We must also rethink our response to the challenge of AI. While restrictions may have their place, outright bans are short-sighted.

Pupils need to learn how to read, summarise and critically evaluate extended texts; they need to develop the capacity to construct balanced and persuasive arguments in writing. The curriculum must continue to require these disciplines. At the same time, it must also include tasks in which pupils are permitted to use AI, but that are carefully structured so that it enhances, rather than replaces their thinking. Tasks that require pupils to use AI to refine, extend and improve the quality of their work, not simply as a means to avoid the cognitive heavy lifting upon which meaningful learning depends.