Gambling Ads vs AI for Kids: Which One Is Actually Harder to Protect Them From?
The bus shelter on Parramatta Road had been wrapped in a Novig ad for about two weeks before my nine-year-old noticed it. He pointed at the screen while we waited for the light to change—some footballer grinning under a headline about five-second deposits. “Why are they giving away money?” he asked. “Can you do that on your phone?”
I gave him the standard brush-off. “It’s not real money, mate. It’s just advertising.” But I watched him read the rest of it anyway, his eyes tracing the odds and the neon colours, and I felt that familiar unease settle in my stomach.
The thing is, I’d spent the previous three weeks worrying about a completely different threat. His school had introduced a “digital literacy” unit that included ChatGPT demonstrations. He came home talking about asking an AI to write him a story about a dinosaur detective. I was concerned about that too. But the more I sat with both worries, the more I realised they weren’t the same kind of problem at all.
The Ads Are Already Regulated (Sort Of)
Here’s what surprised me when I actually looked into it: gambling advertising in NSW is heavily restricted. It can’t be shown near schools. It can’t run during programs aimed at children. There are rules about what the ads can say. There are even time-of-day restrictions.
That doesn’t mean the system works perfectly. The bus shelter outside my son’s school remained controversial for a month before the council complained. Multiple studies show kids see far more gambling content than the regulations intend—through sports coverage, social media, and even video game sponsorship. The ads are designed to feel normal, familiar, part of the texture of Saturday sport and Thursday-night footy.
But at least there’s a conversation happening. At least there are rules that exist, even when nobody enforces them properly. When my son asked about Novig, I could name the problem. I could say, “That’s gambling. It’s a game designed to take money from people. It’s not okay for kids.” I had language for it. We have shared language for it, as a society.
AI Is a Different Kind of Invisible
With AI, I don’t have that language. Nobody does.
My son’s school had a session about how AI tools work, and they gave the kids a worksheet about checking whether information online is reliable. It was well-meaning. It taught them to ask “Is this source trustworthy?” But it didn’t address the bigger problem: an AI conversation feels like a relationship. It feels like talking to someone who understands you. An AI chatbot doesn’t have a commercial motive that you can easily explain—or a regulatory framework you can point to and say, “That’s why this is restricted.”
And the speed of it terrifies me more than any bus shelter ad, because the ads are static. They’re fixed. I can walk my son past a betting ad and deliver a 90-second conversation about why gambling is harmful. But an AI tool grows, adapts, and responds in real-time to whatever my child types into it. I can’t pre-explain every possible interaction. I can’t prepare him for an algorithm that might tell him something upsetting, or confusing, or—maybe worse—something deliberated and gentle and encouraging that he then takes as truth.
I remember a specific afternoon that crystallised this. He was doing a homework project on endangered animals. Instead of using a school-approved resource, he asked an AI chatbot to tell him about the northern hairy-nosed wombat. It gave him a detailed, confident answer that featured not one but several incorrect statistics about current population numbers. When he showed me, I flagged it. But he shrugged. “It sounds right though.”
That’s the difference I keep coming back to. A gambling ad has a visible agenda: it wants you to put money on a game. You can teach a child to recognise that agenda. It’s a social skill, and we’ve been building it for decades. An AI tool’s agenda is much harder to name. It wants to be useful. It wants to be used. It wants to keep the conversation going. The harm isn’t always in what it says—it’s in the trust it builds by sounding reasonable.
What Actually Works in Real Life
After the wombat incident, I didn’t ban AI from his homework routine. That would have been like banning the internet—it’s everywhere now, and pretending you can keep it out just makes it more interesting. Instead, we made a small rule that’s been genuinely useful.
We call it “the second source rule.” Anything an AI tool tells him needs to be checked against another source—a book, a verified website, even a YouTube clip from an actual wildlife organisation. He can still ask the AI questions, but the answers become starting points for the real task: finding out the truth elsewhere.
That’s concrete and it’s worked so far. But I’m honest with myself that it’s a stopgap. The rule works because he’s nine and still thinks I’m a reasonable authority. I know he won’t always share his AI conversations with me. There’s a national road safety ad that says “every parent thinks it won’t happen to them.” I feel that way about the AI issue. I’m doing the basics well, but I’m also aware that the technology is changing faster than my ability to develop reasonable rules for it.
The Novig ad, at least, I have handled. He knows gambling is harmful. He knows the ads are trying to look friendly and fun and they’re not. His teacher covered it again in class after a school session on financial literacy, reinforcing what I’d said. The message is consistent—from me, from school, from public health campaigns. The ad itself is a villain we can both agree on.
What Keeps Me Awake
I don’t think the question “which is worse?” has a useful answer, honestly. The gambling ad is worse for the specific problem it creates—kids learning to see gambling as normal. That’s contained and well-documented. The AI is worse for the uncertainty around it. We don’t know what’s accumulating. We don’t know if there’s a version of this that creates real developmental problems we won’t see for another decade.
My son will be exposed to both. He’s already exposed to both. In the case of the Sydney ads, he’s learning to recognise influences and resist them—an old-fashioned but reliable skill. In the case of AI, I’m less certain. Some days I’m calm about it. Some days I wonder if I’m parenting in a fog, using tools designed for an older version of reality.
I keep the morning routine fairly screen-free. I still read him chapters from actual books. I still answer questions about the wombat population by pulling up a conservation trust website and reading out loud. None of this is profound. None of it solves the big structural questions. But it’s something.
The Novig ad came down off Parramatta Road last week. I was grateful, then realised a different one would probably go up somewhere else soon. We learn to live alongside things we can’t completely control. We make second-source rules. We have conversations about trust and money and who benefits when a child believes what a screen tells them.
It’s not a quick answer. It’s not even a complete one. But it’s the parenting I’ve got for now.
Please indicate: Thinking In Educating » Gambling Ads vs AI for Kids: Which One Is Actually Harder to Protect Them From