What Singapore’s MOE Curriculum Doesn’t Teach About AI
By the time your child finishes Primary 6, they will have spent roughly 6,000 hours in classrooms. Of those hours, about 10 will have been dedicated to coding and emerging technologies. Under the new AI for Fun modules rolling out from 2025, they may get an additional 5 to 10 hours of generative AI exposure if their school opts in.
That is not nothing. But it is not enough either – and most parents do not realise how big the gap is until it is too late to close easily.
This is not a criticism of the Ministry of Education. Singapore’s approach to digital literacy is one of the more thoughtful in the region. The point of this article is simpler: to help parents understand exactly what their child is getting from the national curriculum, where it stops, and what that means for kids growing up in a world where AI fluency will matter as much as literacy and numeracy.
What the MOE Curriculum Actually Covers
Let’s start with what is being taught, because credit is due.
Singapore’s Code for Fun (CFF) programme has been mandatory for all upper primary students since 2020. Jointly run by the Ministry of Education and IMDA, it gives every Primary 4 to 6 student 10 hours of structured learning across:
- Computational thinking — breaking problems into steps a computer can follow
- Block-based coding — typically through Scratch or similar platforms
- Digital making — using microcontrollers like the M5GO or micro:bit
- Introduction to AI — basic concepts of how machines learn
From 2025, the new AI for Fun (AFF) modules add another 5 to 10 hours of optional content for primary students, covering generative AI and smart robots. Secondary students whose schools opt in receive an additional 10 hours covering prompt engineering and AI-driven design thinking.
On top of CFF and AFF, AI is woven into other subjects through the Singapore Student Learning Space (SLS), which now integrates AI-powered tools. Students learn about ethical use, data verification, and how to evaluate AI-generated information across English, Math, and Science contexts. Some secondary schools also offer Computing as an O-Level or Computing+ subject for students who want to go deeper.
For roughly 50,000 students reached every year, this is real exposure to real concepts. It is enough to demystify the technology and plant a seed of interest. The question is whether a seed is enough.
Where the Curriculum Stops Short
Here is what the official numbers do not tell you.
1. The Total Hours Are Lower Than Most Parents Realise
Combine the maximum amounts a student could receive – 10 hours of CFF, 10 hours of AFF, and the digital literacy threaded through other subjects, and you arrive at roughly 25 to 30 dedicated hours of coding and AI learning across an entire primary school career. That is less time than most students spend on a single term of piano lessons.
For comparison, a child enrolled in a weekly 1-on-1 coding class for two years receives over 100 hours of focused instruction, building real projects with feedback every step of the way. The difference is not marginal. It is the difference between knowing what coding is and being able to do it.
2. The Programme Is Optional in Secondary School
Code for Fun is mandatory at upper primary, but only at upper primary. Once your child enters secondary school, whether they continue learning to code depends entirely on whether their school chooses to offer the programme. As of 2023, only about two-thirds of secondary schools opted in.
For students whose passion was sparked at Primary 4, the transition to secondary school can feel like the door closing right when they were ready to walk through it. Unless they pursue Computing as an elective subject — which is itself only available in select schools — formal exposure tapers off precisely when their cognitive readiness for deeper concepts is peaking.
3. The Curriculum Is Built for Breadth, Not Depth
A 10-hour module has to serve every student in the cohort, regardless of prior exposure or interest. That means the pace is set for the median learner, the projects are scoped to be completed in a single term, and the assessment focuses on participation rather than mastery.
This is the right design choice for a national curriculum. It is the wrong design for a child who shows real aptitude. A student who could be writing functional Python by Primary 5 instead spends those 10 hours dragging Scratch blocks they have already outgrown, while a student who needs more support gets the same pace as everyone else. Neither is well served.
4. Tool Exposure Is Not the Same as AI Fluency
The new AI for Fun modules are a meaningful step forward, but it is worth being precise about what 5 to 10 hours of generative AI exposure produces. Students learn what prompts are. They learn that AI can hallucinate. They build a small project. That is genuine literacy.
It is not, however, fluency. Fluency means being able to break a real problem into AI-assisted sub-tasks, evaluate model outputs critically, iterate on prompts to improve quality, and integrate AI into a personal workflow as a thinking partner. That kind of fluency takes time, repetition, and feedback — none of which a short module can provide.
5. The Curriculum Cannot Move as Fast as the Field
National curricula are designed to be stable. They are reviewed in multi-year cycles, approved through committees, and rolled out across hundreds of schools simultaneously. This is a feature, not a bug, when it comes to teaching long-stable subjects like algebra or grammar.
It is a structural problem when the subject itself doubles in capability every 18 months. A Primary 5 student in 2026 is being introduced to generative AI tools that did not exist when the AFF curriculum was being designed. By the time today’s Primary 5 students reach university, the tools they used in school will be three or four generations behind. What stays relevant is not the specific tool but the underlying habit of building, breaking, and rebuilding things with technology — which requires more practice than the curriculum can provide.
Why These Gaps Matter More Than They Used To
A decade ago, most parents could reasonably assume that any gaps in the school curriculum would be filled by the workplace, university, or self-directed learning later in life. Coding was a specialist career path, not a general literacy. AI was something other people built.
That assumption no longer holds.
The students sitting in Primary 4 today will graduate university into a workforce where AI tools are as fundamental as email. Universities, including NUS and NTU, have already begun integrating AI literacy expectations into their entry profiles. Direct School Admission programmes increasingly favour students who can demonstrate genuine technology projects, not just exposure. The international schools — UWC, Tanglin Trust, ACS International — that many Singapore parents benchmark against build coding and AI fluency into their curricula far more deeply than the national pathway does.
The gap, in other words, is not just about whether your child becomes a software engineer. It is about whether they enter their teenage years as a confident creator with technology, or as a passive consumer of tools other people built.
What Parents Can Do to Close the Gap
The honest answer is that closing the gap takes time, not money. A child who codes for an hour a week for two years will be transformed. A child who attends a one-week holiday camp will be entertained.
A few things to look for if you are evaluating options:
- Consistency over intensity. Weekly 1-on-1 sessions over a year produce far better outcomes than a packed holiday programme, because real coding skill is built through cycles of attempting, failing, and iterating with feedback.
- A real progression path. Look for a structure that takes a child from block-based coding through to text-based languages like Python, then on to web development, AI, and project work. A curriculum that loops on Scratch indefinitely is a warning sign.
- Personalisation to your child’s interests. Children who learn to code through projects they actually care about — a Minecraft mod, a game, a website for their hobby — retain the skill. Children who code through generic worksheets do not.
- Real outputs your child can show. A portfolio of finished projects becomes useful for school applications, scholarships, and competitions. It also builds the habit of finishing things, which is itself a life skill.
This is, broadly, the philosophy behind how AlphaGen Learning structures its 1-on-1 programmes for ages 7 to 16. The Sparks track for younger children focuses on creative confidence with block-based tools and simple AI concepts. The Builders track for ages 10 to 12 introduces Python, web development, and machine learning through projects students choose themselves. The Innovators track for teens goes deeper into real programming languages, AI model building, and the kind of portfolio work that supports university and DSA applications.
But the framework matters more than the provider. Whether you choose us, another school, or a self-directed path with a parent who codes, the principles are the same: depth over breadth, projects over worksheets, consistency over crash courses.
The Bottom Line
Singapore’s MOE curriculum gives your child a respectable introduction to coding and AI. It is more than most countries provide, and it has improved meaningfully with the addition of AI for Fun in 2025.
It is also, by design, an introduction. Twenty hours of structured exposure across a school career is enough to know what AI is. It is not enough to be fluent in it. The children who will move through their teenage years and into adulthood as confident builders with technology will be the ones whose parents recognised that gap early and chose to close it.
The earlier you start, the smaller the gap to close. The longer you wait, the harder it gets.
Frequently Asked Questions
How many hours of coding does my child get in Singapore schools? Upper primary students receive 10 hours of mandatory Code for Fun lessons across Primary 4 to 6. From 2025, an additional 5 to 10 hours of optional AI for Fun content is available. Secondary school participation is optional and varies by school.
Is Code for Fun enough to prepare my child for an AI-driven future? Code for Fun provides a strong introduction to computational thinking, but 10 to 20 hours of national curriculum content is unlikely to produce genuine fluency. Children who develop real coding and AI skills generally supplement school exposure with consistent weekly practice over multiple years.
At what age should my child start coding outside of school? Most children can begin block-based coding meaningfully from age 7. Children with prior exposure to Scratch in school often benefit from moving to text-based languages like Python around age 10 to 12.
Does coding outside school help with DSA or school applications? A portfolio of completed coding and AI projects can strengthen Direct School Admission applications, particularly for talent areas including STEM, technology, and innovation. Schools increasingly value demonstrated project work over generic enrichment attendance.
Curious how your child’s current skills compare to what they could be building? Sign up for a free 60-minute trial class — our instructors will assess where your child is and show you what a personalised pathway could look like.
