Neither banning nor ignoring works well. A small number of household rules and one reliable test — can they do a similar question tomorrow? — covers most of it.
Why banning rarely works
These tools are freely available and children use them regardless. A ban mostly moves the use out of sight, which removes your ability to see how it is being used.
It also forgoes the genuine benefits, which exist. The aim is supervised use rather than no use.
Why ignoring it is worse
Unmonitored, most children drift towards asking for complete solutions, because that finishes the homework fastest and the incentive is obvious.
The damage is invisible for months. Homework is completed, marks in class are unaffected, and the gap only surfaces in an exam.
Rules that are actually enforceable
Attempt first, tool second. Hints before solutions. Explain anything the tool produced. Use it in a shared space rather than alone in a bedroom.
Enforceability matters more than comprehensiveness. Three rules that are followed beat ten that are not.
Four rules worth setting:
- Attempt the question before asking anything
- Ask for hints, not solutions
- Be able to explain any answer produced
- Use it where you can be seen, not alone
The test that tells you the truth
Give a similar question the next day without the tool. That single check reveals whether the homework produced learning, and no amount of discussion substitutes for it.
If they cannot do it, the issue is not the tool but how it was used, and the conversation can be about that rather than about technology in general.
It is worth doing this every couple of weeks rather than only when you are suspicious. Made routine, it stops being an accusation and becomes an ordinary part of how homework is checked — which is both fairer to the child and considerably more likely to catch a genuine problem early, while there is still time to do something about it.
What lessons at Odyssey are actually like
A student on how the lessons work in practice. The clip is from our secondary classes, but the way concepts are explained is the same at every level. More are on Odyssey Math TV.
Where Odyssey fits
We set homework to be attempted independently and review it in the next lesson, which makes it visible fairly quickly when work has been produced rather than understood.
We teach maths only, in small classes at Hexacube, 160 Changi Road #01-02, about three minutes from Eunos MRT. See the primary and secondary programmes, our MOE-aligned curriculum, schedules and fees, or the tutors who teach. Free video solutions to past-year papers are on Odyssey Math TV.
Almost every primary topic is scaffolding for a later one, and the scaffolding is invisible until it gives way. That is why our curriculum is written across all six years, Primary 1 to Primary 6, rather than level by level — it lets us see the whole structure a child is standing on, not just the step they are currently on.
If you are unsure whether tuition is the right answer yet, the free trial is a reasonable way to find out without committing to anything. We will give you an honest read on where your child stands and what would actually help. Book a free trial class or WhatsApp +65 8512 6882.
What we describe here comes from teaching this syllabus in our own classes since 2012. It is not a forecast for any individual child: progress is produced jointly, by a student willing to do the work with support at home. We do not promise grades. Topics follow the MOE primary mathematics syllabus — check the current syllabus on the MOE website for the definitive version, and see our curriculum page for how we teach it.
Frequently asked questions
Should I ban AI tools for maths?
Generally not. Banning tends to move use out of sight rather than stop it, and it forgoes the real benefits. Supervised use is a more workable goal.
How do I know whether my child understood the homework?
Give a similar question the next day without the tool. It is the only reliable test.
What if my child says everyone else uses it?
They may well be right. The useful conversation is about how it is used rather than whether.
Is using AI to check answers cheating?
Checking after an attempt is ordinary good practice. Generating work to submit as your own is a different matter, and worth naming as such.
