At the curriculum night in October, I explained our classroom AI policy to a room split roughly in half: families who’d moved from East Asia in the last five years, and families from Europe and North America who’d been at the school longer. The policy was the same document for both groups. I’d written it myself.

About ten minutes in, I noticed the room had diverged. The Western parents were nodding at the part about critical thinking and academic integrity. The Chinese parents were leaning forward when I mentioned that students would be using AI tools regularly in class.

They weren’t having the same conversation. Neither group was wrong. They just weren’t reading the same policy.

A study published this month in Computers & Education explains why. Researchers interviewed parents in China and in Western countries about their attitudes toward generative AI for tutoring. What they found was a clean divergence in how each group thinks about the central question.

Chinese parents see AI adoption as a competitive imperative. Not using AI isn’t a neutral choice — it’s falling behind. If other students are using AI tutoring to get ahead, opting out is a form of disadvantage. They’re worried about being left behind by not adopting.

Western parents are more cautious. They recognize the same technical advantages — personalization, breadth, accessibility — but weigh them against concerns about dependency, academic integrity, and data privacy. They’re worried about what’s lost by adopting.

Both groups are responding rationally to the information they have and the environment they’re thinking about. Chinese parents are thinking about a highly competitive educational system where marginal advantages compound. Western parents are thinking about a child who might stop thinking for themselves, or whose work can’t be verified as their own.

Both groups of parents exist in most international schools. They sit at the same curriculum nights. They read the same emails from the principal. They’re supposed to be governed by the same AI policy.

Most school AI policies are written from one set of concerns.

Western educators are almost universally trained in systems where academic integrity is the primary lens. The instinct is toward caution: spell out what students can and can’t do, create a structure that treats AI as a risk to be managed, emphasize critical thinking as the counterweight to AI dependency.

That’s a coherent policy. It makes sense given the concerns Western parents have.

But it doesn’t speak to what Chinese parents are worried about. A policy that says “students will use AI with guidance and with attention to academic integrity” reads differently if your concern is “my child will fall behind children who have unrestricted access to the best tutoring tools available.” The caution that reads as responsible stewardship to one parent reads as institutional reluctance to another.

Neither parent is wrong about the policy itself. They’re working from different assumptions about what the stakes are.

This is harder at international schools because the divergence isn’t hypothetical — it’s in the building. You can’t write a policy for “parents” as a unified stakeholder group, because the parents in your community don’t share the same concerns.

The study found that both groups agreed on the technical merits of generative AI: it’s personalized, it’s accessible, it knows a lot, it’s available at 2am when a student is stuck. That common ground is real. But agreement on the tool’s capabilities doesn’t resolve the disagreement about whether and how to use it.

That gap is where school AI policies usually fall short. They establish what students can do without explaining the reasoning in terms both parent groups can locate themselves in. A Chinese parent reading “we approach AI use thoughtfully” doesn’t know whether that means “we integrate it fully because the benefits are real” or “we limit it because the risks are real.” A Western parent reading “we prepare students for an AI-integrated future” doesn’t know whether that means “we teach them to use it critically” or “we give them unrestricted access.”

The words are doing too much work, trying to speak to everyone while landing differently for each group.

What the study found that most school communications miss: the Chinese parents’ competitive-pressure concern isn’t irrational or anxious in a way that should be dismissed. It’s grounded in a real observation — that AI proficiency is likely to matter, that early exposure compounds, that educational advantage is built incrementally.

The Western caution is also not irrational. Dependency is a real risk. The question of what students can do without AI is a legitimate one.

These aren’t opposing errors. They’re opposing priorities, and both deserve a direct response in how you talk about AI with families.

Most school AI communications are written as if the parent reading them already agrees with the underlying concerns. “AI use should be transparent and purposeful” — sure, but purposeful toward what end? Toward building proficiency that compounds, or toward preserving the ability to think without it? Both parents might agree with the sentence and mean opposite things by it.

The result is a policy that nobody fully reads themselves into, and both groups sense that. The Western parent wonders if the school is keeping up. The Chinese parent wonders if the school is holding back. Neither asks the question out loud at curriculum night, because the room doesn’t signal that the tension is acknowledged.

When institutions don’t name the divergence, families fill in the gap with their own interpretation of which side the school is really on. And whichever side they think the school favors, they assume the policy is designed for the other group’s comfort, not their child’s needs.

Takeaway for Teachers

Before your next curriculum night or parent communication about AI, add one sentence that names the split explicitly: “We know families come to this conversation with different concerns — some are focused on making sure students have real access to these tools, and others are focused on making sure students can still think and work without them.”

That’s not a policy position. It’s an acknowledgment that the room isn’t unified, and that you see it.

When you signal that you see both orientations, parents stop reading the policy as coded support for the other side. The Chinese parent stops hearing “responsible use = we’re limiting your child’s access.” The Western parent stops hearing “AI-integrated future = anything goes.”

You don’t need a different policy for different families. You need a policy that shows you understand why the same sentence reads differently in different households.

If you’re thinking through how to communicate with families about AI — and what no algorithm can tell you about the students sitting in between — my book goes deeper. The AI Doesn’t Know Your Students is available on Amazon and at shouldiuse.ai/book.

Get the next piece before it’s published. One email, most weeks — what I’m seeing in the classroom, nothing else.

David Jacobson is a high school history teacher. He writes about AI, education, and the messy intersection of the two at shouldiuse.ai.