At some point on almost every school tour now, a teacher opens a laptop, types a prompt, and a chatbot spits out a poem about the ocean in about four seconds. Everyone smiles. It’s a nice moment. It just doesn’t tell you much.
Mei found that out the hard way. She’s a made-up parent, but a familiar one, touring HK international schools with her nine-year-old son. After the third poem demo in a week, she stopped being impressed and started asking different questions. What did the children actually learn? Who checks that they understood it? And where does all their data go?
Those are the questions worth taking with you.
What does “good” AI teaching even look like?
Not a child getting a chatbot to do their homework, that’s for sure. A school that takes AI seriously teaches kids to use it, to doubt it, and to put it down when a task needs their own brain. And it does that inside normal lessons like history, science and English, not in a one-off “tech week.”
There’s a useful yardstick. In 2024, UNESCO published an AI competency framework for students that calls for AI learning objectives to be written into school curricula, so children can use these tools safely and meaningfully. You don’t need to read it cover to cover. Just ask where those kinds of objectives turn up in your child’s year group, and how teachers check the kids have got it.
Mei’s favorite trick was simple. Instead of asking “Do you use AI?” (every school says yes), she asked to see one recent lesson. Say a Year 5 history class used AI to draft a summary of the Opium Wars. What did the pupils hand in? What did they check against a textbook or another source? Could they talk about the topic afterwards with the laptop shut? The schools she liked could show her real pupil work where you could see the child’s own thinking. The ones that said “we’ll get back to you” went into her notebook.
Are the teachers ready for this?
Fair question, and the local research suggests it’s worth asking. Between May and June 2026, the Education University of Hong Kong collected 1,892 valid questionnaires from schools and other education settings. Teachers were broadly positive about AI, but many said they needed more professional guidance and hands-on support. That’s background, not a verdict on any international school. Still, it’s a good reason to dig a little.
“Have your staff been trained?” is too easy to answer with a yes. Mei asked something harder to fake: how did the training change the way you plan a lesson? One teacher told her they now build in a step where pupils fact-check the chatbot’s answer against the class textbook. That’s the kind of concrete answer you want.
It’s also worth asking how AI use grows over the years. What does it look like for a seven-year-old compared with a thirteen-year-old? Is there language support for kids still getting comfortable in English? Does every child take part, or just the ones in the after-school coding club? With little ones especially, ask exactly how the teacher keeps an eye on things.
What are the rules, and who sets them?
Any school using AI should have a written policy, and you’re allowed to see it. The good ones spell out which tools are approved, who supervises them in each year group, and how parents hear about changes. If Stamford American School Hong Kong is on your list, start with its campus and curriculum pages. Then ask for a recent lesson example and the AI rules for your child’s year.
Next, follow the data, because this is where vague answers tend to hide. Do names, photos, voice recordings or pupils’ work ever get uploaded to an AI service? Who can see them? How long are they kept? Are they used to train somebody’s model? And if a tool isn’t right for a younger child, what does that child do instead?
Two recent developments give you some handy language. UNICEF updated its guidance on AI and children in December 2025, and it lists safety, data privacy, fairness, transparency, inclusion and well-being among its requirements for child-centred AI. Those make a neat set of headings for your questions, though remember the guidance doesn’t certify any school’s practice.
Then there’s Hong Kong’s own move. In July 2026, the Privacy Commissioner for Personal Data and the Digital Policy Office launched the Safeguarding Personal Data AI Sandbox. Applications for its first phase came from publicly funded primary and secondary schools, which most international schools aren’t. So the sandbox tells you nothing about how a particular international school protects your child’s information. You’ll need to ask them directly.
Will my child still learn to think for themselves?
Probably the question most parents care about most. Kids are already using AI for schoolwork, with or without permission. So ask how teachers draw the line between help that’s allowed and work a pupil has to do solo. Ask to see how pupils own up to using AI when it is allowed.
If the school offers IB courses, there’s a clear rulebook. The IB’s academic integrity policy treats AI-generated content like any other source. When a task allows it, pupils have to credit it. Passing it off as their own counts as misconduct. Whether AI is allowed on a given assignment is still up to that assignment’s rules, and the school should be able to explain them in plain words.
And one softer question that’s easy to forget: who looks after your child when something’s wrong? UNICEF’s updated guidance now covers AI companions used by children, which is a good nudge to make sure there’s always a real, named adult a pupil can go to. Staff should be able to tell you who that is without hesitating.
Mei’s three-question cheat sheet
By her fourth tour, Mei had boiled everything down to three requests, and she asked the same three at every school.
The first was a recent lesson example. A good answer names a specific lesson and shows what pupils made and what they checked. If it’s vague, push: could the pupils explain the work without the tool?
The second was the written rules. A good answer is an actual policy naming approved tools, supervision arrangements, and data handling for each year group. If it’s fuzzy, ask who keeps it updated and how parents find out when it changes.
The third was how teachers give feedback. A good answer shows how a teacher responds to a draft when AI was allowed and when it wasn’t. If they’re hesitant, ask for an anonymised example of a real comment on a draft.
She kept three columns in her notes: what she’d actually seen, what staff had told her, and what she was still waiting for. A slick demo rarely made it past the second column.
One last tip from her notebook. If a school offers IB courses alongside its own diploma, ask whether the AI rules differ between the two and how teachers explain the difference to pupils. Rules on AI vary a lot between schools, so get the answer for your child’s year group in writing.
Making the final call
Once the tours are done, look at what’s still unanswered and pin down who’s sending the missing policy or lesson example. Then put AI where it belongs, as one factor alongside the commute, the language of instruction and whether your child actually lit up walking around the place.
It really comes down to three things: learning, privacy and independent work. Judge each school on what it can show you, not what it promises, and pick the one that suits your child’s age and needs.
FAQs
Q. Should every pupil use AI?
No. Ask which tasks allow it, which need independent work, and what the alternatives are.
Q. What should parents ask about privacy?
What data gets uploaded, who can see it, how long it’s kept, and whether it trains any models.
Q. Can a school promise AI is safe?
A blanket promise isn’t enough. Ask for the written policy, the supervision setup, and how updates are shared.
Q. How do I compare schools fairly?
Ask each one for a recent lesson, an anonymised piece of feedback, and the rules for your child’s year, then note the gaps.
Related: Human Skills in the AI Era: 6 Skills That Matter Most in 2026
