AI and young children
Commentators on AI generally fail to distinguish between younger and older children. There is now plenty of evidence that too much screen time (computers, smart phones and TV) is damaging to mental health and make it less likely that children will learn to concentrate and become proficient readers. The 2025 PISA scores and commentary show this. Schools and parents should strictly limit screen time for all children and young people. Primary-age children need have nothing to do with AI.
The whole curriculum
All secondary pupils should be given in-depth lessons about AI, a tool which will be useful throughout their lives.
Memorisation of useful knowledge is one important purpose of education. Having useful knowledge in your long-term memory, such as speaking a language or remembering a poem, adds to human enjoyment. It is not enough that information is to be found through AI.
An AI machine may well know more about any subject than a human but that does not detract from the pleasure all humans derive from possession of knowledge in their brains which can be called upon throughout their lives. Apart from anything, knowledge can give us discernment, including the ability to question the output of an AI machine.
There is an argument for adjusting the curriculum a bit so that human beings (pupils), and especially teenagers, learn to do things that machines are less likely to do and things which bring enjoyment to the individual. People are different from machines and we need to emphasise people’s humanity and individuality explicitly.
Here are some examples of valuable activities which go beyond ten GCSEs and three A-levels:
1. Opportunities to do something original which reflects the interests of the teenager, like the EPQ.
2. Extend beyond activities of the head (academic learning) to activities of the hand (making things) and the heart (social activity). Make these hand and heart activities compulsory – they are valuable skills and part of what make us human.
3. Create systems which give every pupil an adult mentor they can relate to. We want teenagers to seek advice and consolation from people, not chatbots. In boarding schools this person is often the house parent. In some schools he or she is the personal tutor. Let the personal tutor teach something of their enthusiasms once a week for a couple of hours.
4. Develop their ability to speak well, insisting that all pupils reach a good level. See the School 21 oracy programme. Most pupils can be taught to speak effectively, in both formal and informal settings.
5. Give teenagers choices where possible, such as which sports they take part in, or writing an article of their choice for a school magazine. They welcome choice.
6. Allow pupils to take worthwhile risks – such as starting a Young Enterprise company or going on a Duke of Edinburgh expedition or directing a play.
Teaching methods
AI can be used to develop individual, personalised teaching as an alternative to a teacher teaching 30 children simultaneously. This will grow in significance but is limited by the fact that too much screen-time is damaging, and the personalised approach still requires a teacher to assess what is going on with each pupil.
Pupils’ work
Extended writing coursework is almost bound to be corrupted by AI use. This is why Ofqual is going to squeeze extended writing out of A-levels. Universities are increasingly replacing extended writing with invigilated exams and orals.
Quite apart from cheating, the problem with using AI is that it prevents pupils thinking for themselves or thinking at all. We believe that humans still need to know things, and use of AI will make it much less likely humans will think or learn.
Use of AI can act to limit reading from books. In that respect it diminishes the ability of pupils to read and make their own notes from texts, which are valuable skills. We must not allow AI to kill off reading and writing for pleasure.
So we should limit pupils’ use of AI.
Marking
AI can be successfully used for marking. It is quick and can give detailed feedback to the teacher and pupil. It can generate a summary of what was done well and what less well. It can be more accurate than a teacher. Eg SURPASS.
However, Ofqual will not permit AI marking of public exams. It is not as reliable as tech companies claim. It is very difficult to get AI to explain a mark it has given.
Marking is something which creates a relationship between the pupil and teacher, if done in person. Marking is used to motivate pupils, but this is less easy if a robot has done it.
Lesson planning
With the usual caveats, AI can plan lessons well and is excellent for generating knowledge/notes/ideas. It is good for image creation.
However, using AI for all lesson planning strips teachers of their creativity and autonomy. An AI lesson plan does not cater for the knowledge, experience and personality of the teacher, which are all-important.
Finally
AI has more potential for some school subjects than others. All subjects are different.
AI is good for report-writing and creation of policies. It can reduce teacher workload.
By Professor Barnaby Lenon, Dean of the Faculty of Education at The University of Buckingham
Leave a Reply