HK Just Bet HK$2 Billion on AI in Schools. Here's What That Actually Means for Your Child.

HK$2 billion is a large number. Policy announcements attached to large numbers tend to produce two responses: enthusiastic press coverage in the short term, and slow, uneven implementation over the following years that rarely matches the headline ambition.

I don't say this to be cynical. I say it because I've watched education policy in Hong Kong for a long time, and the gap between what gets announced and what changes in actual classrooms is consistently wider than anyone in the announcement room acknowledges. Parents who make decisions based on the announcement rather than the implementation reality are working with incomplete information.

So let me try to give you the more granular picture — what this money is likely to actually do, what it won't do, and what you should be asking your child's school.

What the government has announced

The HK$2 billion AI in education fund covers several categories: infrastructure (devices, connectivity, server capacity for AI tools in schools), curriculum development (new AI literacy components across subjects), teacher training, and piloting of AI-assisted teaching tools in selected schools.

The direction is clear and, in my view, broadly correct. Hong Kong has decided to integrate AI into core school curricula rather than restrict it. While other jurisdictions are still debating whether to allow AI tools in classrooms, Hong Kong is asking how to teach students to use them well. That's the right strategic bet. The question is execution.

AI literacy: primary vs secondary

The policy distinguishes between primary and secondary AI education, and the distinction matters.

At primary level, the vision is "AI literacy" — a term that can mean many things but in practice refers to conceptual understanding of what AI systems are, how they work in broad terms, and how to interact with them critically. Think: understanding that AI tools make predictions based on patterns in training data, not understanding. Understanding that AI can be wrong and that outputs need to be verified. Understanding how recommendation algorithms shape what you see. These are not coding skills. They're interpretive and critical skills, and they're genuinely important.

What this means in a P4 classroom is still being worked out. Curriculum frameworks have been drafted. How those frameworks translate into lesson plans, and how those lesson plans are delivered by teachers who may have limited AI background themselves, is the gap that HK$2 billion has to bridge.

At secondary level, the ambition is more substantial: AI as a subject within the computing and STEM curriculum, including practical components where students work with AI tools to solve real problems. The stronger secondary schools with established ICT departments are positioned to implement this well. The schools without dedicated computing teachers — and there are many — face a genuine challenge.

The 80% problem

A recent survey found that around 80% of Hong Kong teachers already use AI tools in some form. This figure is often cited as evidence that Hong Kong's teacher readiness is strong. I read it differently.

Using AI tools personally — using ChatGPT to draft a letter, using an AI grading assistant for routine marking — is not the same as being able to teach students to use AI thoughtfully. The skills required to use a tool and the skills required to teach critical engagement with that tool are different. A teacher who uses AI to save administrative time has learned efficiency. A teacher who can help a student evaluate an AI output, understand where it might be wrong, decide when to trust it and when not to — that teacher has pedagogical AI literacy, which is a more demanding capability.

Teacher training is the component of this investment I'll be watching most closely. If the training is primarily about adopting specific tools — here's how to use this AI marking platform, here's how to set up this writing assistant — the professional development will produce technically capable but pedagogically shallow results. What teachers actually need is the kind of conceptual grounding that lets them engage critically with AI tools they haven't seen before, because the specific tools will change continuously.

What isn't changing

The DSE. Not yet.

The core high-stakes examination sits largely outside the AI integration push. This is, in some ways, a structural contradiction at the heart of the policy: schools are being asked to teach students to work effectively with AI tools while preparing them for examinations that test their ability to perform without those tools.

This isn't an unsolvable contradiction — the skills are genuinely complementary if taught well. A student who understands AI deeply enough to use it critically is also developing the analytical thinking the DSE rewards. But it creates an allocation problem for schools. Time spent on AI literacy is time not spent on DSE preparation in the conventional sense. Schools with strong DSE track records will be cautious about rebalancing too aggressively toward new curriculum components whose examination outcomes are uncertain.

The practical result, for the next two to three years: AI integration will happen faster in co-curricular and project-based contexts than in core examination preparation. Your child's computing class will look different before their Chinese history class does.

What a parent should ask their school right now

Not "does your school have an AI strategy?" Every school has a strategy document now. Strategy documents are cheap.

Ask instead:

What specific changes will I see in my child's curriculum this year? If the answer is vague — "we're monitoring developments," "we'll be integrating AI across subjects" — the implementation is not yet real. Ask for the specific subjects, specific year groups, specific timeline.

How are your teachers being trained? Ask about the duration, provider, and content of teacher professional development on AI. A one-day workshop is not teacher training. Sustained, subject-specific development over a term is closer to what meaningful change requires.

How does the school distinguish between using AI as a learning tool and using AI instead of learning? This is the question that reveals whether the school has done genuine pedagogical thinking or whether it's just adding AI tools to existing workflows. The answer should be specific. It should reference assessment design, not just policy statements.

What happens when a student uses AI in a way that's not disclosed? The school's answer to this reveals how seriously they've thought through academic integrity in the new environment. Vague answers here are a warning sign.

My actual assessment

The HK$2 billion signals something important: a governmental decision that AI integration in education is not a threat to manage but a capability to develop. That commitment, sustained over years, can produce meaningful change.

But policy timelines are measured in years and implementation quality is uneven. Your child is in school now. The schools that will serve your child well in this transition are the ones that have already started — that have teachers who are genuinely engaging with AI pedagogy, not just waiting for the official curriculum rollout. Those schools exist. They're not defined by budget or prestige. They're defined by whether the teachers in the classroom are thinking seriously about what good AI-era education looks like.

Ask the question. The answer will tell you more than any government announcement.