A mother messaged Empire AI last week.

“I’ve banned ChatGPT in our house. My daughter is only nine.”

She was asked one question: have you asked her friends?

She came back two days later. Her daughter had been using AI for three months.

At a friend’s house. On the school iPad. In group chats with classmates.

The ban hadn’t kept AI out of her daughter’s life. It had only kept the mother out of the conversation.

CNA interviewed 15 Singapore parents about MOE rolling out AI to Primary 4. Two-thirds were not opposed to AI, they were opposed to uncertainty.

They wanted vetted tools, transparent rules, and a seat at the table.

That isn’t anti-tech. That’s good parenting.

But here’s what got buried in the headline:

The AI your child uses at school is not the AI you’re afraid of.

What you’re imagining vs. what’s actually happening

When you read “AI in Primary 4,” your mind probably jumps to your nine-year-old typing “write my essay” into ChatGPT at midnight, alone in their room.

Here’s what MOE is actually rolling out:

● AI delivered through the Student Learning Space (SLS): the same regulated national platform Singapore schools have used since 2018.
● The classroom AI is called LEA (Learning Assistant). It does not give answers. It asks questions back.
● Before students touch it, they complete self-paced modules on what AI is, how it works, and how to use it responsibly.
● Every interaction is teacher-supervised.

That is not ChatGPT in a bedroom at 11 pm. That is structured pedagogy with guardrails.

Two completely different things keep getting conflated. They shouldn’t be.

The fears nobody is talking openly about

Most parents aren’t really afraid of AI itself. They’re afraid of:

● Falling behind their own child.
● Not knowing the right answer when their child asks them about it.
● Their child getting hooked before anyone explains the rules.
● Brain damage. Real, neurological, “is this making my kid dumber” worry.
● Looking irresponsible if they say yes too early.
● Looking reckless if they say no too late.
● Being the parent who got it wrong while everyone else got it right.

CNA didn’t quite name those. So let us answer them.

Fear #1: “Will AI rot my child’s brain?”

The research is real and worth taking seriously. SMU’s Asst Prof Jacqueline Ho is right: at primary age, neural pathways are still forming, and over-reliance on AI for thinking-tasks can impede literacy, numeracy, and critical thinking.

But notice the word: over-reliance.

A calculator doesn’t ruin a child’s maths if the child still learns the maths. AI doesn’t ruin a child’s thinking if the child is still being taught to think.

The risk isn’t AI. The risk is unsupervised, unstructured, undirected AI use, which is exactly what happens when parents ban it at home and the child finds it through friends.

Fear #2: “Isn’t Primary 4 too early?”

Probably yes, for typing prompts at 2am, alone, in a bedroom.

Probably no, for sitting next to a parent or teacher and asking “hmm, is that actually true? How would we check?“

Same technology. Two completely different relationships with it.

The age isn’t the problem. The framing is.

Fear #3: “I don’t even understand AI. How can I guide them?”

This is one of the most common fears parents quietly express.

The answer is simpler than many expect:

You don’t need to be an expert. You need to be curious in front of your child.

Sit beside them. Open ChatGPT, Claude, or Gemini. Ask it something you already know the answer to. Watch what it gets right and what it gets wrong. Then, out loud:

“Hmm. It got that wrong. Why do you think it did that?”

That single moment teaches your child more than any school policy ever will. It teaches them AI is a tool, not an oracle. It teaches them that questioning the tool is the job of the human using it.

It also surprises parents. AI is both more capable and more limited than they imagined. That surprise replaces fear with understanding.

The 3 camps every Singapore parent falls into

Camp #1: Ban it.

Won’t work. Your child will find it through friends, school iPads, group chats. Your ban only removes you from the conversation.

Camp #2: Hand it over.

Won’t work either. Unstructured AI use at this age is exactly what the research warns about. This is how brains get under-trained.

Camp #3: Engage with intention.

Sit beside them. Use it together. Ask questions of it. Disagree with what it says. Build something with it. Together.

Camp 3 is the only one that scales with your child. Because the AI in their life isn’t going anywhere and the one they’ll be using in five years will run a third of the economy by the time they finish O-Levels.

The protection isn’t keeping AI away from your child. The protection is a child who knows how to think with AI without losing themselves to it.

What Parents Need To Hear

Your concern is valid. The research on over-reliance is real. Trust that instinct.

Your ban won’t hold. Not because your child is defiant because the technology is already woven into their world, their friends’ phones, their school platforms, every device they will ever own.

The best protection is a prepared child. Not a child who has never seen AI. A child who has been taught to ask: Who made this? Why might it be wrong? What do I bring to this that AI cannot?

You don’t have to figure this out alone. The conversation MOE is trying to start, however imperfectly, is the right one. So is the conversation with your child at the dinner table tonight.

The line CNA’s article ended on:

Empire AI strongly agrees with this principle.

What we do at Empire AI is not teach children to use AI instead of thinking. We teach them to think through AI, to direct it, question it, build with it, and recognise when not to use it.

The children who will thrive in an AI-shaped Singapore are not the ones who used it the earliest.

They are the ones who understood it the deepest.

And that always starts with a human asking a good question.

Where do you stand?

Are you in the ban it camp, the let-them-use-it camp, or somewhere in between? Have you actually sat down and used an AI tool with your child yet?

The conversation continues at Empire AI.

Register for a Free Trial AI or Robotics Class

Our MOE-registered educators work with children aged 4 to 19. Parents are welcome to sit in on a free trial class alongside their child and explore AI together in a guided, structured environment. Please complete the form below and our team will be in touch with you shortly.