A polite machine

1The first time I asked a chatbot to outline an essay on harbour reclamation, it answered in thirty seconds and called me “of course”. The outline was tidy: three benefits, two drawbacks, a conclusion that recommended “further study”. It sounded like a student who had memorised a template and was afraid to offend anyone, including the sea.

2Teachers now talk about AI the way they once talked about Wikipedia: first as a scandal, then as a tool, then as a fact. Some schools ban it in exams and ignore it in homework. Others ask students to paste the machine draft and mark, in a different colour, what they changed. A few pretend the software does not exist, which is a form of nostalgia.

3The danger is not that the machine is brilliant. It is that it is polite and average. It will not tell you that your example is from 2011. It will not admit it has invented a statistic. It will not get bored and therefore will not stop. A human teacher’s most useful sentence is still “this paragraph does no work”. A chatbot rarely says that about its own output.

4Used honestly, the software can be a sparring partner: ask it for the opposite view, then attack the opposite view. Used lazily, it becomes a ghostwriter whose English is smoother than yours and whose thinking is thinner. Markers are not stupid. They have read a hundred essays that all recommend further study in the same calm voice.

5What should count as cheating is not the existence of a tool but the disappearance of a process. If you can explain your outline without looking at the screen, the machine has been a calculator. If you cannot, it has been a substitute. Schools that only police the exam hall and not the fortnight before it are guarding the wrong door.

6I still open the chatbot when I am stuck. I also close it before I write the first real sentence. The harbour does not need another polite outline. It needs a student who has stood at the railing, smelled the mud, and decided what they actually think.