Raising a Kid in the AI Era

This episode is a bit different. I’ve had my wife Tavani join me for this one, and I’m glad I did because she sees all of this from a completely different angle. She’s not in cyber security. She’s not thinking about AI professionally every day. She’s just a parent watching the same future roll in, and honestly that perspective cut through a lot of my overthinking.

The topic: raising kids in the AI era. It keeps me up at night more than any ransomware stat or zero-day. My son is four years old, and I already feel completely underprepared for the world he’s going to grow up inside.

He’s Four, and It’s Already Here

Carson is four. He’s not asking ChatGPT anything. He’s not using AI tools. But we already are, on his behalf. Snotty nose in the middle of the night? Rather than calling the doctor or Googling it the old way, we’re typing symptoms into Claude. Sleep problems, talking in his sleep, weird rashes. The reflex now is to ask AI before anything else.

And I don’t think that’s a bad thing, exactly. But it’s worth noticing. The habit we’re building right now, reaching for AI before we reach for a person, is exactly what he’s going to pick up by watching us. Kids don’t learn from what you tell them. They learn from what they see you do at the kitchen bench at 10pm.

That’s the part that sits with me. He already knows how to swipe left and right on an iPad. He watched me enter my PIN once and memorised it. Figured out how to unlock it himself. He’s four. And AI is only going to accelerate everything around him from the moment he actually starts caring about it.

The Skills That Matter More, Not Less

Something that came up in our conversation that I can’t stop thinking about is this: the real problem isn’t that AI will replace his job one day. The problem is that if he grows up leaning on it for every decision, he won’t be able to think for himself when there’s no internet connection, no device, no AI in the room. He’ll just be stuck.

I keep coming back to the GPS analogy. I am absolutely terrible with directions now. Tavani gives me grief about it constantly. What are you going to do if Google Maps isn’t there? And honestly, I don’t have a great answer. That’s one tool, applied to one narrow problem, and it’s already cost me something real. Now imagine that across everything a kid faces as they grow up.

The spell check parallel is another one that hit me during the conversation. When computers got spell check, we all got a bit lazy about spelling. When calculators showed up, mental arithmetic started to fade. Each of those was one small thing. AI is not one small thing. It’s everything, all at once, and it’s going to be so baked into the tools around him that opting out won’t really be possible by the time he’s old enough to make the choice.

So the skills I want him to have before AI gets to help with them, like reading carefully, writing by hand, sitting with a hard problem long enough to actually feel it, those matter more now, not less. Not because AI is bad, but because you need the foundation to know when the AI is wrong. I’ve talked about this before in the context of Kubernetes: if I hadn’t spent years actually learning Docker and container networking, I wouldn’t have had the instinct to push back when Claude Code was leading me down a path that didn’t feel right. The knowledge is what lets you challenge the tool. And you can’t shortcut building that knowledge by asking the tool itself.

The Dark Side Schools Won’t Teach Fast Enough

This is probably where our conversation got most serious.

The deepfake problem. The synthetic video problem. The sextortion problem.

I work in cyber security. I see the phishing attempts, the voice cloning, the AI-generated impersonation attacks. I give presentations on ransomware statistics that apparently scare my wife quite a bit. Her words, not mine. But connecting those things to what a high school environment looks like in five or ten years, that hit differently. That’s not abstract anymore. That’s my kids world.

There’s a story from Tavani’s primary school that stuck with me. Someone printed a fake letter, put her name on it, and the damage it did was real even though the technology was primitive. Now imagine the same social cruelty with tools that can generate convincing video of a real person doing something they never did, in minutes, for free. The technology that I struggle to identify as fake in a professional context is going to be in the hands of fifteen-year-olds. And that’s already starting to happen.

I don’t believe schools are going to get this right quickly. They’re going to focus on the productivity angle of AI, the curriculum integration, the lesson planning tools. The dark side of it, the part that will actually affect his mental health and his social world, is going to lag behind. We’ve already seen this play out with social media. South Australia tried a ban on social media for under-16s and kids found workarounds within days. A ban was never going to be the answer. Education was. And I suspect the same thing happens here.

Our parents had no framework for the chat rooms and the internet risks we navigated as teenagers. They weren’t in those spaces. They didn’t know what we had access to. We kind of figured it out ourselves, and sometimes it cost us. The difference now is that Tavani and I have lived through it. We know what these tools can do at a professional level before they’ve fully hit the general population. That awareness, uncomfortable as it is, is probably the only real advantage we have as parents.

Modelling the Behaviour, Not Just Stating It

Right, so here’s the thing I keep landing on: you can’t tell your kid one thing and do another. You can’t sit there on your phone reaching for ChatGPT every time a question comes up and then lecture him about the importance of thinking for himself. He’s not listening to the lecture. He’s watching the phone.

We can’t be using AI constantly and then tell him he’s not supposed to use it. Same as screen time. You can’t watch YouTube all evening and then tell him his thirty minutes is up.

So there’s a real discipline piece here that applies to me, honestly more than it applies to him right now. I’ve caught myself reaching for AI to draft emails at work and I don’t love what that’s doing to my own voice. Tavani still prefers to write the email herself and then have AI clean up the grammar, because otherwise it doesn’t sound like her. Use it to help you be you, not to be you instead of you. That’s what I want him to see modelled, not just hear explained.

I’m not fully optimistic that we’ll get the balance right as parents. But I’m pretty clear on what getting it wrong looks like, and that’s a kid who’s fast, capable-looking, and completely lost the moment the tool goes away.

Carson is probably going to be fine. He’s a good kid, and he’s going to grow up and figure out the world the way kids always do. In twenty years he’ll probably be explaining AI to us in ways we can’t currently imagine. But fine isn’t the same as thoughtful. And the part I have a say in, right now, at four years old when he’s watching me unlock my phone and wondering what I’m typing, that part is about whether he grows up seeing me think carefully, or watching me skip the thinking entirely.

That’s the bit that keeps me up at night. And honestly, I don’t have a confident answer. But I reckon asking the question is a better start than not.

As always, keep on learning.