What the EDB's 2024 Generative AI Guidance Actually Says (Translated for Non-Educators)

The Education Bureau has issued several documents addressing generative AI in schools, but the language of official education policy is not written for parents. Terms like "scaffolded inquiry," "learning objectives alignment," and "responsible deployment frameworks" are clear to educators and opaque to everyone else.

Let me translate the key points, fill in what the guidance doesn't say, and give you the practical implications.

What the EDB guidance actually says

The core messages of the EDB's generative AI guidance can be summarised in a few plain-language statements.

Generative AI is here and schools should engage with it, not ignore it. This is perhaps the most significant shift from initial instinctive reactions in some schools towards banning all AI tools. The EDB's position is that students will live and work in a world with AI, and schools have a responsibility to prepare them. A blanket ban is not the recommended approach.

Schools have discretion in how they apply this. The EDB is not mandating a specific AI curriculum or specific approved tools. Individual schools can decide whether and how to use generative AI tools in teaching and assessment. This means the actual experience varies enormously — some schools are moving quickly, some are not moving at all, both under the same official guidance.

Academic integrity must be maintained. AI assistance is not the same as academic dishonesty in all cases, but students should not submit AI-generated work as their own independent thinking. Schools are expected to set clear expectations and explain these to students.

Teachers need training and support. The guidance acknowledges that most teachers have not had formal preparation for teaching in an AI-enhanced environment and calls for professional development. This is important because it implies the EDB recognises the implementation gap.

Students should develop AI literacy as a skill. Using AI tools critically, understanding their limitations, evaluating their outputs — these are positioned as learnable capabilities, not threats to be avoided.

What the guidance doesn't say

Some things that parents might expect to be in official guidance are conspicuously absent.

There is no approved list of AI tools. The EDB has not said "use these and not those." This means schools are making their own decisions about tools, without consistent quality standards or vetting.

There is no standard assessment policy. Whether a student declaring AI use on an assignment is commended or penalised, and how much, is left to individual schools and teachers. This inconsistency is a genuine problem for students who receive different signals in different classes.

There is no defined level of AI literacy expected at each grade. There's general language about developing AI literacy, but no equivalent of "by P6 students should be able to X" for AI capabilities. This makes it hard to evaluate whether your child's school is adequately covering this area.

There is no clear guidance on data privacy with AI tools. Several AI platforms popular with students collect significant data. The EDB guidance doesn't specifically address whether schools should assess privacy implications before using or recommending tools.

What this means in practice

The consequence of discretionary, high-level guidance is wide variance. In practical terms, this means:

Your child's experience of generative AI in education depends almost entirely on their individual school's choices. Two students at different schools in the same district may have radically different experiences — one receiving sophisticated, integrated AI literacy education and the other receiving nothing or a blanket ban.

Teachers within the same school may apply different standards. One teacher may permit AI brainstorming assistance; another in the same department may prohibit any AI engagement. Without school-level policy, this is hard for students to navigate.

The EDB's general intentions are sound but the gap between intention and implementation is real. Good policy at the guidance level doesn't automatically produce good practice at the classroom level.

What to ask your child's school

Given all this, here are specific questions worth asking at the next school open day or parent meeting.

"Does the school have a written policy on student use of generative AI tools? Where can I find it?"

"Are students taught about the limitations and risks of AI-generated content, not just how to use AI tools?"

"Is AI literacy addressed in the curriculum formally, or only through individual teacher initiative?"

"When students use AI assistance for an assignment, what are they required to disclose? What are the consequences of non-disclosure?"

Clear, specific answers to these questions will tell you far more about the school's actual approach than general statements about embracing technology.

The parental role in the gap

Where school guidance is unclear or inconsistent, the clearest place for students to receive principled guidance on AI use is at home. The principles I'd suggest:

Be honest about what you used. If AI helped you produce something, say so. This is a version of citation, not confession.

Understand what you submit. Whatever tool helped you produce work, you must be able to explain and extend it. If you can't, you haven't learned.

Question the output. AI is often right and sometimes confidently wrong. Treat it as a first draft or a thinking partner, not a final authority.

These principles are good guidance now and will remain good guidance as the tools continue to evolve.