An Assembly on Artificial Intelligence from Alex Grant
September 23, 2026
Deep in the summer holiday, my dishwasher broke down. It started beeping maniacally, relentlessly and making a deep, unhealthy gurgling sound. We are dependent on technology in so many aspects of our lives, it is frustrating when that dependency is abruptly cut off. An August of suds and rubber gloves was an unwelcome prospect.
I have little skill to fix mechanical things. Very little. Faced with this problem, I thought I’d turn to someone - or something - in my household that might help: my own intelligence fell short so I thought I’d see if Artificial Intelligence could assist me in my predicament. Could one machine fix the other? I fed the symptoms of my dishwasher’s distress into the Large Language Model, Google Gemini. ‘What should I do?’ I asked it, pathetically. It quickly diagnosed the problem: water was trapped in an inaccessible part of the machine. Now, I do own an assortment of under-used screwdrivers; with a flash of inspiration, I suggested that I could unscrew some prominent screws at the bottom of the dishwasher. ‘No!’. Gemini advised me. Sensibly, it told me to not do that under any circumstances or I would break the thing forever. Screwdriver hanging limply in my hand, my exchange with Gemini continued…
You will know Artificial Intelligence is the subject of much debate. Some claim, optimistically, that it will revolutionise human life for the better, solving difficult problems and removing drudgery. On the other hand, some claim unrestrained AI development could lead to the extinction of humanity. That would, at least, solve the dishwashing problem.
For those of us in schools, the emergence of generative AI poses a challenge to the purpose of education; it can now do many of the tasks that have, until now, been the preserve of humans. Put simply: AI can – efficiently – find any information, read and summarise a text, write fluent and polished prose, offer instantaneous translation, complete every part of a mathematical or scientific problem, and generate music or art in any genre. Is it, therefore, still important to learn to do these things for ourselves.
Before Easter, I asked senior year groups for their views on AI. I am grateful to the 400 in U5 and above who responded with eloquence, supported by their experiences. Whilst many noted positives - that AI can bring efficiencies – it was notable that the majority expressed concern, frustration or even anger about the impact that AI may continue to have.
Some cited environmental concerns, or problems of biased, deepfake or hallucinated information. Many noted that AI can compromise academic integrity and enable cheating. A majority expressed concern that dependence on AI may result in people losing the ability to think critically or independently.
This view is increasingly supported by hard evidence. A study by MIT in the USA last year indicated that dependence on AI for essay writing resulted in students having a much shallower understanding of the subject, weakening their critical thinking and formation of long-term memory. A recent Chinese study of 27,000 students found the exam performance of those who used AI for their regular work was worse than those who did not.
Many of you said you are confused about what is permitted and that the school has not given sufficient guidance. That’s a fair point. The answer is not simple, but I can set out some broad principles.
Firstly, in the age of AI, the cultivation of your intelligence and creativity remains an intrinsic good. Intellectual and creative effort has value. Even if AI can do a task, we should be interested in developing our own learning for its own sake; to encounter astonishing new things that change us as human beings.
To let you into a secret: when I set a history essay to a class, it isn’t because I actually need 25 copies of that essay. It is because the process of producing the essay develops your learning. Efficiency in production is not the goal of education. Learning requires your brain to work hard and you can’t bypass that difficulty or offload it onto someone or something else. It would be analogous to repeatedly driving a car for 26 miles in the expectation that you will be fit and prepared to run a marathon.
Secondly, if AI is used, it should be to support learning, not undermine it. There are AI tools that can be powerful in helping to develop thinking and your teachers may introduce these to you. But their use should not circumvent intellectual effort that is a necessary part of learning.
Thirdly, we recognise the growing importance of AI literacy; as the technology becomes embedded into our lives, it is essential that we understand its capabilities and limitations, as well as ethical questions that are raised in a world where AI may become ubiquitous. Mr Fisher has done pioneering work on this with his Fifth Form Elective course and I look forward to working with Lawrence Doe on his red gown project to develop how we do this further.
Finally, AI must not diminish our humanity or our relationships in this community. The ‘real’ matters and it was clear in your feedback that many of you are repelled by AI-generated material that is synthetic or impersonal. In particular, you felt strongly that if you put time, effort and love into something, that you want it recognised and considered by a teacher, not assessed by a machine.
As with all technology, AI should remain a tool in service to our needs. Dependency on technology for intellectual tasks – tasks much greater or more complex than washing dishes - runs the risk that we become incapable.
So, what was the resolution of my dialogue with Google Gemini? It concluded – firmly - that I should step away from the dishwasher, lay down my screwdriver, and call out a mechanic; a real person with the knowledge and skills – the human intelligence - required. When he arrived, he sorted the problem within 15 minutes. I could return – sheepishly - to my dependency on the machine.











