When ChatGPT first exploded onto the scene, I was genuinely excited. Not the cautious, wait-and-see kind of excited — I mean, I stayed up late reading about it, thinking about what it could mean for how I teach, how I give feedback, how I plan. This felt like something real.
And then I went to school the next day and had absolutely no idea what to do with that excitement.
Whom was I supposed to talk to? Was anyone else at my school even thinking about this? Was I allowed to use it? Should I be using it with students? I didn’t know where to start, and I didn’t know whom to follow — inside my school or anywhere else. So for a while, I did what a lot of teachers did: I experimented quietly on my own, figuring it out as I went, hoping I wasn’t accidentally doing something I shouldn’t be.
I don’t think that’s a unique story. I think it’s the story of most teachers over the last couple of years — including, I’d guess, a lot of people reading this right now. And if you’re teaching at an international school, particularly in China, that sense of isolation can be even more pronounced. You’re already operating far from the professional networks and conversations happening in your home country, and the AI tools everyone is talking about may not even be accessible where you are.
So this is the article I wish someone had handed me when I was staying up late reading about ChatGPT and wondering what came next. It won’t answer every question — honestly, nobody can do that yet — but it’ll give you a clearer picture of the landscape, the real risks worth caring about, and what responsible AI use actually looks like in practice.
First, the Honest State of Student Protections
Here’s something that surprised me when I dug into it: the two main federal laws protecting student data in the U.S. — FERPA (passed in 1974) and COPPA (1998) — were written decades before generative AI existed. They weren’t designed for a world where a student can paste their essay into a chatbot, and that text potentially ends up in a company’s training data.
The good news is that states are catching up fast. As of 2025, more than half of U.S. state departments of education have issued AI guidance for K-12 schools, and 53 bills were proposed across 21 states in just the last legislative session. Updated COPPA rules also took effect in June 2025, adding stronger restrictions on how companies can share children’s data with third parties.
But here’s the thing — if you’re teaching at an international school in China, U.S. federal law doesn’t follow you there. You’re operating in a completely different regulatory environment, and that changes the calculus significantly.
A Different Playing Field: International Schools in China
Teaching at an international school in China means navigating two overlapping realities that most professional development resources simply don’t address.
On one side, many international schools — particularly those following an American, British, or IB curriculum — maintain their own institutional data privacy policies modeled on Western frameworks. Your school may have agreements with ed-tech vendors that mirror FERPA-style protections, and your students’ families likely enrolled expecting those standards to be upheld.
On the other side, you’re operating within China’s own regulatory environment. China’s Personal Information Protection Law (PIPL), which came into full effect in 2021, is in some ways more stringent than U.S. law — it restricts how personal data can be transferred outside of China and requires explicit consent for data collection. What this means practically is that using a U.S.-based AI tool with student data could potentially put that data on servers outside of China, raising compliance questions under both Chinese law and your school’s own policies.
Then there’s the access layer. Many of the AI tools most popular among Western educators — ChatGPT, Google’s Gemini, and similar tools — are blocked by the Great Firewall. This creates a dynamic that doesn’t get talked about enough: teachers and students may be using VPNs to access these tools, which puts them in a legally gray area under Chinese internet regulations and almost certainly outside the scope of any vendor agreement your school has negotiated.
What fills that gap? Chinese domestic AI tools — Baidu’s ERNIE Bot, Alibaba’s Tongyi Qianwen, and others — are fully accessible and growing rapidly in capability. But these tools operate under Chinese data governance rules, meaning data collected by these platforms is subject to Chinese government access in ways that are very different from what Western families typically expect or consent to.
This isn’t a reason to panic or avoid AI altogether — it’s a reason to be clear-eyed and intentional.
The Risks That Actually Matter Day-to-Day
Wherever you’re teaching, these three realities are worth understanding before you use any AI tool in or near your classroom:
Consumer tools and education tools are genuinely different. Products built specifically for schools — like ChatGPT Edu — don’t train their models on student data. Many free, consumer-facing AI tools do. In China, this distinction applies to both Western tools accessed via VPN and domestic Chinese platforms, each carrying their own data handling implications.
You’d be surprised what counts as student data. Most teachers think of student data as grades or transcripts. But a name combined with a school ID, a disability status, or even a recognizable writing sample can constitute protected personally identifiable information. For international school families in China — many of whom are expats with heightened sensitivity around data sovereignty — this matters even more.
The jurisdiction question is genuinely unsettled. If a student at your Beijing international school submits work through a U.S.-based AI platform, where does that data live? Who has access to it? Your school’s legal team may not have a clear answer. That ambiguity is itself a risk worth flagging.
So What Does Responsible Use Actually Look Like?
Let me get practical, because this is where the conversation usually gets too vague. Here are a few real scenarios, with the international context built in.
Scenario 1: Writing feedback at scale
You have 120 essays due Friday — a very familiar situation, whether you’re in Houston or Hangzhou. You’re tempted to paste them into an AI tool to generate feedback faster. Here’s how to do this responsibly: strip out any identifying information before you prompt the AI. Instead of “Here’s Marcus’s LEQ on the Silk Road,” write “Here’s a student LEQ on the Silk Road — evaluate it against this rubric.” You get the efficiency, the student’s data stays protected, and the feedback you generate is still yours to personalize before it reaches them.
For teachers in China specifically: use only tools that your school has explicitly approved, and be wary of using any platform accessed via VPN for anything involving student work. The convenience isn’t worth the legal and ethical exposure.
Scenario 2: Lesson planning and differentiation
This is the lowest-risk, highest-reward use of AI for teachers, and it applies equally in any country. Asking an AI to help you build a scaffolded reading on the Mongol Empire for struggling readers involves no student data at all — you’re using AI as a planning partner. For teachers in China, this is also where domestic AI tools can actually shine: tools like ERNIE Bot are capable of generating lesson content in both English and Chinese, which can be genuinely useful for differentiation in a multilingual classroom. Just apply the same critical eye to the output that you would with any AI tool.
Scenario 3: A student wants to use AI for a project
This is where the conversation gets more nuanced, and in an international school context, it gets even more so. Students in China may have very different levels of access depending on whether they’re boarding students, day students, or working from home on a family network. Be explicit about which tools are approved by your school and accessible within the regulatory environment you’re operating in. And consider making the AI interaction itself part of the assignment — have them document their prompts, reflect on what the AI got right or wrong, and show their own thinking alongside it. That’s good pedagogy anywhere, and it also naturally surfaces the access and tool questions that students in China are already navigating.
Scenario 4: AI-generated feedback on assessments
Several states and international accrediting bodies are beginning to restrict fully automated AI decision-making around student assessment. AI can flag patterns, suggest scores, or draft comments — but the professional judgment has to be yours. Your knowledge of your students — their context, their growth, their specific struggles — is something no AI has access to. That’s especially true in international schools, where students are often navigating language barriers, cultural transitions, and family circumstances that significantly shape their performance.
The Habits That Will Keep You on Solid Ground
Stick to school-approved tools, full stop. In any context, your school’s approved tool list exists for a reason. In China, this matters even more — approved tools have presumably been vetted against both your institution’s policies and the local regulatory environment. Tools accessed via VPN, however convenient, fall outside that protection.
Default to anonymous prompts. Any time you’re putting student work into an AI tool, remove the identifying details first. It takes ten seconds and eliminates the biggest risk, regardless of what country you’re in.
Ask your administration what policies are actually in place. You might be surprised how many schools — international ones especially — haven’t fully worked this out yet. Asking the question isn’t a sign that you’re behind. It’s a sign that you’re paying attention. And it might be the nudge your school needs to get serious about it.
Be upfront with students and families. In international school communities, families have often made very deliberate choices about where their children’s data lives and who has access to it. A brief, honest explanation of what tools you’re using and why goes a long way. And if you don’t have clear answers yet, say so — that transparency builds trust rather than undermining it.
Find your people. This is the one I wish someone had told me earlier. You don’t have to figure this out alone, and you don’t have to wait for your school to hand you a professional development session. LinkedIn is genuinely one of the best places right now to find educators who are thinking seriously about AI — follow people who are asking hard questions, not just celebrating every new tool. Look for the international educator community specifically; it’s active, it’s thoughtful, and a lot of those teachers are navigating the exact same tensions you are.
The Bottom Line
The pressure to adopt AI so our students stay “competitive” is real and growing — and it’s felt just as acutely in Shanghai and Beijing as it is in any American classroom. But the most important thing we can model for our students isn’t how to use the latest tool. It’s how to engage with powerful technology thoughtfully, critically, and responsibly.
I started excited but directionless, experimenting quietly and hoping for the best. I’d like to think I’ve gotten more intentional since then. Not because I have all the answers, but because I’ve started asking better questions — about the tools I use, the data I handle, and the example I’m setting for the students in front of me every day.
Wherever you are in that journey — whether you’re staying up late reading about AI and wondering what to do next, or you’re already deep in the weeds — you’re not alone. And the conversation is just getting started.
What are you doing in your classroom to navigate AI responsibly? I’d genuinely love to hear from educators at every stage, and from every corner of the world.
