Generative AI and Academic Integrity: The Policy Every HK School Needs but Few Have

Here is a simple test. Find your child's student handbook or school code of conduct. Search for the word "artificial intelligence," "ChatGPT," "generative AI," or "AI tools." In my experience, roughly 70-80% of Hong Kong schools currently return zero results.

This is a problem. Not because AI policy is the most important thing schools should be doing, but because in the absence of clear policy, students are making decisions without guidance, teachers are enforcing standards inconsistently, and parents are in the dark about what's expected.

Why clear policy is hard to write

I want to be fair to schools here before I criticise them. Writing AI policy is genuinely difficult, for several reasons.

First, the technology is changing faster than policy processes. A policy written in September may be outdated by February when a new model capability changes what's possible.

Second, the right answer differs by subject and task type. Using AI to generate a Chinese composition is very different from using AI to research background information for a science project. A blanket policy ("AI is allowed" or "AI is not allowed") doesn't capture these distinctions.

Third, there's genuine disagreement in the education community about where the lines should be. Some teachers believe that using AI for writing assistance is simply how writing will work in the future and schools should embrace it. Others believe that writing ability is a foundational skill that must be developed without AI assistance, particularly for younger students. Both positions have merit.

What good policy looks like

A well-designed school AI policy doesn't try to resolve these debates universally. Instead, it provides a framework that allows different rules in different contexts, with transparency about those rules.

The framework I'd advocate for has several components.

Declaration requirements. Students should indicate when they have used AI assistance on an assignment, just as they indicate when they've used other sources. Not as confession, but as citation. "I used [tool] to [purpose]" tells the teacher something useful and normalises honest disclosure.

Task-specific rules. Each assessment should specify whether AI assistance is permitted and, if so, for which purposes. "You may use AI tools for brainstorming and outlining; the written expression must be your own." Or: "This assessment requires independent work only." Or: "AI tools are permitted throughout; your reflection log should describe how you used them." These distinctions matter and should be explicit.

Competency-based verification. For assignments where significant AI assistance is possible, schools should build in components that require demonstration of understanding — a short oral component, a follow-up question, a reflection on the process. This isn't about catching cheating; it's about ensuring the assignment still achieves its learning purpose.

Consequences that fit the behaviour. Using AI to understand a concept better and then writing your own work is different from submitting AI output without engagement. Punishing both identically sends the wrong signal.

What to do while your school figures this out

If your child's school doesn't have a clear AI policy, or has a blanket ban that treats all AI use identically, there are things you can do at home to ensure they're on solid ethical footing.

Set your own household rules that are more nuanced than the school's. "You can use AI to help you understand things you're stuck on, but what you hand in must be your own thinking, and you must be able to explain everything in it." This is a defensible position that most teachers would endorse even if they haven't articulated it in policy.

Teach your child to ask teachers when in doubt. "Is it okay to use AI for this?" is a completely reasonable question. A teacher who can't answer it has a policy gap. Most teachers will appreciate the question and give a considered answer that your child can follow.

Teach the principle behind the rules, not just the rules. The reason AI-generated work is problematic isn't because it breaks rules — it's because submitting work that doesn't represent your thinking means you haven't done the thinking. The learning has been bypassed. This principle applies whether the school has articulated a rule or not.

The school's responsibility

I want to say clearly: the burden here should not primarily fall on students or parents. Schools have a responsibility to provide clear, current, specific guidance on acceptable AI use. Expecting students to navigate this without guidance, and then disciplining them for doing so "incorrectly," is not fair.

If your child's school hasn't addressed this, it's worth raising as a parent. Ask at the next Parent-Teacher Association meeting: "What is the school's policy on AI tool use for homework and assessments?" If the answer is vague, suggest that a clear policy would be helpful.

Schools respond to parental interest. Collective pressure from engaged parents can accelerate policy development in ways that individual teacher advocacy sometimes can't.

The bigger picture

Academic integrity has always been about learning to think and taking ownership of that thinking. AI introduces new tools and new temptations, but it doesn't change the fundamental purpose of education. A student who has genuinely understood, genuinely thought, and genuinely produced something that represents their thinking has not violated integrity, regardless of what tools they used along the way.

The policy question is ultimately a question about how we structure learning in an environment where powerful tools are available. Get the structure right and AI becomes an accelerator for genuine education. Get it wrong, and we've just made it easier to go through school without learning anything.