Educating boys to think in the age of AI
There is an assumption developing around artificial intelligence (AI) that Christ’s College needs to think deeply about. Because AI is transforming the world beyond the school gates, the argument goes, it must inevitably transform what happens within them. If students are going to use AI in the workplace, surely schools have a responsibility to embrace it too. I think this argument requires serious reflection.
AI is an extraordinary technology. It will undoubtedly change industries, workplaces, and many aspects of our lives. However, is AI designed to best support the purpose of schools?
Christ’s College has a very particular purpose – we exist to educate boys to think. Before we rush to embrace a technology capable of doing so much for our students, we should ask whether this technology enhances our boys' capacity to think or could it diminish it?
To understand this we need to look at our first principles. As a school, we want our boys to learn how to reason, question, communicate, solve problems, exercise judgement, and develop ideas of their own. That purpose has survived every technological change of the past and, in my view, it should remain at the centre of education in the age of AI. We become competent at tasks often through repetition. We learn to write by writing. We learn to problem solve by calculating. We develop arguments by constructing and testing arguments. We learn to read difficult material by reading complex literature. There are no shortcuts to developing these capabilities. Yet shortcuts are precisely what AI is exceptional at providing.
If a student is asked to write an essay and AI helps him formulate the argument, structure the response, and improve the language, we might end up with a better essay. However, have we developed a better writer? If AI can summarise a difficult text in seconds, the student may understand its basic content more quickly, but has he become a better reader? These distinctions matter because education is not primarily about producing work, It is about developing the person producing the work. Perhaps, as Dr Tom Hawkins pointed out to me, the simplest analogy is physical exercise. If we want to become stronger, we go to the gym and place our muscles under stress. There needs to be resistance. If somebody else lifts the weights for us, the weights still move, but we do not become stronger. Intellectual development is not so different.
A student sitting in front of a blank page, uncertain how to begin an essay, is experiencing intellectual resistance. So, too, is the student who cannot immediately solve an equation or the student who is struggling to understand Shakespeare. Our instinct as educators should not be to remove that resistance. Sometimes, the resistance is the education. I worry that AI is beginning to steal that struggle from our boys. Every individual intervention can appear helpful, yet, collectively, they risk removing precisely the repeated intellectual effort through which competence is developed.
This is particularly important during adolescence. School is not the workplace for our boys. A 45-year-old using AI to write a routine report more efficiently is drawing on decades of accumulated knowledge, vocabulary, experience, and judgement. A 14-year-old is still developing knowledge and skills. We should, therefore, be very cautious about importing the language of workplace productivity into education. Efficiency is valuable in business because the outcome is generally the objective. Education is different because the process is the objective.
Schools are not supposed to be efficient in everything we do. Reading an entire novel when a summary is available is inefficient. Discussing an idea with a teacher for half an hour when an answer could be generated in seconds is inefficient. However, in education, inefficiency can be immensely productive. The time, repetition, frustration, and effort involved are often where the learning occurs. AI reverses this relationship. It is overwhelmingly focused on the outcome. A polished answer produced with minimal intellectual effort may be an impressive product, but it can represent very little learning.
There is another concern, and it applies as much to teachers as it does to students. AI promises to make teachers more efficient. Again, that sounds entirely desirable. However, we should ask what might be lost when we automate apparently mundane parts of teaching. Consider school reports. A colleague recently pointed out to me that writing hundreds of reports is time-consuming, and AI can undoubtedly make the process faster. However, when a teacher sits down to write about our boys, they are required to stop and think about that individual. How has he progressed? Where is he struggling? What motivates him? What does he need from me next? Over time, there is a risk that teachers spend less time thinking deeply about the boys they teach.
Education is fundamentally relational. The quality of the relationship between teachers and our boys remains one of the most important influences within a school. Great teachers know their students. They develop this knowledge through thousands of small human interactions, many of which are inherently inefficient. The more we insert technology between teacher and student, the more vigilant we need to be about what it might displace.
AI has largely been developed to make human beings more productive: to generate content, synthesise information, answer questions, and automate cognitive tasks. Those are powerful capabilities in an economy seeking greater productivity. However, a school is trying to do something quite different. We are deliberately asking inexperienced young men to perform cognitive tasks so that, through performing them repeatedly, they become more capable.
The question, therefore, is not simply whether AI can improve a student's work. Of course it can. The more important question is whether AI can improve a student's work while simultaneously making the student less capable of producing that work independently. I think that risk is real.
There is also something profoundly important about achievement that we should be careful not to lose. The deepest satisfaction in learning rarely comes from receiving an answer. It comes from reaching one. These are not peripheral benefits of education. They are among its most important outcomes.
I believe schools should resist the assumption that widespread adoption of AI is inevitable or desirable. The world our boys inherit will be faster, more automated, and increasingly saturated with artificial intelligence. Perhaps that makes it more, rather than less, important that school remains a place where they are required to think for themselves. A place where they read the book rather than the summary, and struggle with the equation rather than request the solution.
Most importantly, school must remain a deeply human place. Education happens through relationships. A teacher who knows a boy, challenges him, and believes in him offers something no artificial intelligence can replicate.
There will undoubtedly be pressure on Christ’s College to demonstrate that we are at the forefront of AI adoption. We should be careful not to confuse technological adoption with educational progress. If the purpose of education is to develop independent thinkers, then we should be very wary of a technology whose great attraction is its ability to think on our behalf.
Our boys have an entire lifetime ahead of them in which machines will make things easier. School may be one of the last places where we can deliberately allow some things to remain difficult. We should not be too quick to take that gift away from them.
Read on
