The honest answer: it depends on their age, and nobody explains that properly
Your kid mentions they used ChatGPT for something. Maybe you knew. Maybe you didn't. Either way you're now doing the mental maths on whether that was fine, with very little to go on.
Here's what most advice gets wrong. It treats a seven year old and a sixteen year old as the same problem. They're not even close.
The bit most parents haven't been told
ChatGPT requires users to be at least 13. The parental controls OpenAI released only work on accounts aged 13 to 18.
So if your child is in primary or intermediate, they are not a supported user of the product at all. There are no parental controls available to you, because from the platform's point of view the account shouldn't exist.
That's not a scare tactic. It's just a fact that changes what you should be doing.
If your child is under 13
You are the safety layer. There isn't another one.
- Use it together, on your account. Not theirs. Same room, screen visible.
- Treat it like the internet in 1999. Curiosity is good. Unsupervised curiosity is where it goes sideways.
- Ask what they asked it. Make it a normal conversation, not an interrogation. You want them telling you things.
- Watch for the confident wrong answer. This is the big one at primary age. AI states nonsense with total authority, and a nine year old has no reason to doubt it.
If your child is 13 to 18
You have actual tools. Link your account to theirs and you can:
- Set quiet hours, so it won't respond late at night
- Turn off voice mode
- Turn off memory and chat history
- Limit image generation
- Turn Study Mode on by default
What you cannot do is read their conversations. That's a deliberate design choice, not an oversight. You'll only hear from OpenAI in limited situations where their systems detect a serious safety concern.
Worth knowing: either of you can unlink at any time, and you'll be notified if they do.
The three questions actually worth asking
Forget "is AI safe" as a question. It's too big to answer. Try these instead.
- Is it age appropriate? Not just the content, but the framing. A tool built for adults will answer an eleven year old like an adult.
- Is it truthful? Assume it isn't, until checked. Teach the check, not the fear.
- Is it replacing thinking? This is the one that matters most and gets discussed least.
Where New Zealand is actually at
Around 60% of young people here already use generative AI for schoolwork. Roughly 69% of NZ teachers use AI weekly.
The Ministry of Education has published guidance for schools, but it's high level, and the Education Minister has publicly acknowledged it could be better. Schools are largely working this out as they go.
Which means the gap is being filled at home, by parents, mostly without support. That's the actual problem, and it isn't your fault.
What I'd do this week
One conversation, not a lockdown. Ask what they've used, what it got wrong, and what they'd ask it if they could ask anything. You'll learn more in ten minutes than any monitoring app will tell you.
Then find out what your school's AI policy says. Most parents have never asked, and every school is required to have a position.
.svg)
