My kid started using ChatGPT for homework and I didn't know whether to be proud or horrified

My daughter is eleven. She found ChatGPT through a classmate, and within two weeks she was using it to check her English compositions. I discovered this when I saw the browser history and felt what I can only describe as a confusing simultaneous rush of "that's actually quite resourceful" and "this is definitely cheating."

Both reactions were partially right. That's what made it hard.

Let me tell you what she was using it for, because the details matter. She wasn't asking it to write her compositions for her. She was writing the composition herself and then asking ChatGPT to "check if it sounds natural." She was using it the way I use spell check, or the way I once used a dictionary. She would read the suggestions, make some changes, ignore others. She had some aesthetic opinion about her own writing that she was defending even against AI feedback.

This is not what I expected. I expected the laziest possible use case: paste the assignment question, copy the output, submit. That's what I would have done at eleven if the technology existed and I am not proud of this about myself.

What she was doing was more interesting. She was using it as a proofreading tool and, in the process, having opinions about her own work. That last part — the opinions — was the thing I hadn't expected.

The problem: her school's position on AI tools is essentially a vague prohibition on "using technology to do your work for you," which is not specific enough to be enforceable and not nuanced enough to address what she was actually doing. Her teacher, when I asked, admitted the school hadn't formally discussed AI policy in primary. This is the situation in most Hong Kong primary schools right now: the children are ahead of the institution.

I had two choices. I could ban it, which would send it underground and mean I lost visibility over how she was using it. Or I could engage with it honestly, set some parameters, and make it part of a real conversation about what homework is for and what learning actually means.

I chose the second, with some nervousness.

Here are the parameters we agreed on. She writes everything herself first, before consulting any AI tool. She reads AI suggestions and decides for herself whether to accept them. She cannot use AI for Chinese homework, because the whole point of Chinese writing in P5 is developing her own facility with the language, and AI scaffolding would hollow that out entirely. She can use it for English as a proofreading layer, not a drafting tool. She tells me what she used it for on any given piece of work.

We have had more interesting conversations about writing since this began than in all the years before. Why did she reject a suggestion? What did she think was wrong with her original phrasing? What does "sounds natural" even mean, and who gets to decide? These are genuinely good questions about language that she is engaging with at eleven because a chatbot gave her feedback she partially agreed with and partially didn't.

I won't pretend I'm comfortable with all of this. There is a version of this that goes wrong — a version in which she becomes dependent on AI feedback, loses confidence in her own judgment, stops developing independent proofreading instincts because she's outsourced them. I watch for these signs.

There is also a version that goes right — a version in which she grows up knowing how to use these tools intelligently, critically, as aids rather than replacements, in the same way that people who learned to use calculators well can also do arithmetic and people who learned to use GPS can also read maps.

We are, I suspect, at the beginning of the longest homework negotiation in history. Every generation of parents has had to work out what the new tool means for learning. We had calculators. Our parents had televisions. Our children have this.

What I'm most grateful for is that she told me about it — or rather, that the browser history told me and she didn't deny it and the conversation that followed was actually honest. Some families are navigating this in secret. We're navigating it in the open, which is the only way I know how to do it.

Ask your children what they're using. The answer might surprise you.