By David Jacobson | March 24, 2026
Last week, one of my AP World History students turned in a source analysis that cited a YouTube video as primary evidence for the fall of the Mongol Empire. The video was slickly produced, confidently narrated, and almost entirely wrong. When I asked her why she trusted it, she shrugged and said, “It had two million views.”
That moment stuck with me — not because it was unusual, but because it’s becoming the norm. We’re living in what Colleen Kenny calls “content chaos”: the collapse of shared frameworks for evaluating information. And if you’re a teacher in 2026, you already know that AI didn’t create this problem — but it’s accelerating it in ways we can’t afford to ignore.
The World Is Moving Fast. Schools Need to Keep Up.
Here in Shanghai, I watch two education systems respond to AI in real time — and the contrast is striking.
China’s Ministry of Education has rolled out a national AI curriculum across primary and secondary schools. Third-graders learn AI fundamentals. Fourth-graders tackle data and coding. By fifth grade, students are working with intelligent agents and algorithms. The ambition is enormous: AI education in all schools by 2030, and fully integrated into textbooks and exams by 2035.
Meanwhile, international schools in China are charting their own path. As AmCham China recently reported, some schools are threading AI into the curriculum from Early Years through Senior School — starting with Bee Bots and Dot and Dash for little ones, moving through LEGO SPIKE Prime and micro:bit, and ending with VEX V5 Robotics. Others, like ISA Wuhan International, are taking a more cautious approach, emphasizing the “human touch” even as they acknowledge AI’s promise.
Neither approach is wrong. But both are grappling with the same question: what do students actually need to know?
It’s Not About the Tools. It’s About the Thinking.
I’ll be direct: the most urgent skill we can teach right now isn’t prompt engineering or Python. It’s the ability to evaluate what’s real.
Kenny’s argument resonates deeply with anyone teaching in an international school context. She points out that trust in U.S. media has dropped to 28%, with similar declines across democracies globally. Algorithmic curation has replaced the old media gatekeepers with something worse: personalized information silos where citizens can’t agree on basic facts. And now generative AI has made it so that, as Kenny writes, “the average person can no longer tell the difference between real video and AI-generated video.”
For my students — kids from dozens of countries, consuming media in multiple languages, navigating both Chinese and Western information ecosystems — this isn’t abstract. It’s Tuesday.
A recent Ed Week survey found that nearly 70% of middle and high school students are worried that AI is eroding their critical thinking skills. That number has climbed sharply in the past year alone. And 70% of teachers share that concern. The students are telling us something. We should listen.
What I’m Actually Doing in My Classroom
Here’s what this looks like in practice — in an AP World History class in Shanghai, where I use AI regularly and intentionally.
I use Claude as a thinking partner for lesson design, feedback drafts, and assessment building. When I’m writing feedback on a student’s Document-Based Question essay, AI helps me generate a first draft that I then revise with my knowledge of that particular student — their growth trajectory, their confidence level, what they need to hear next. The AI doesn’t replace my judgment. It gives me a starting point so I can spend more time on the human part.
But here’s the thing I keep coming back to: I also teach my students about AI, not just with it. We do source analysis exercises where students have to evaluate AI-generated historical summaries against actual primary documents. We talk about why a confidently written paragraph can be completely wrong. We discuss who built these tools, whose data trained them, and whose perspectives get amplified or erased.
Robin Lake at the Center on Reinventing Public Education has documented how different countries are approaching this challenge. Singapore is making AI teacher training mandatory by 2026. South Korea is building adaptive homework systems. Finland’s “AI in Learning” initiative prioritizes equitable access. Each approach reflects different values and priorities — but the through-line is clear: every serious education system in the world is reckoning with AI right now.
The Ethics Piece Isn’t Optional
China’s own guidelines offer a useful framework here. The Ministry of Education prohibits primary school students from independently using open-ended content generation tools, and bans teachers from using generative AI as a substitute for core teaching. That’s not anti-technology — it’s a recognition that guardrails matter, especially for young learners.
The European Commission’s updated guidelines on ethical AI in education reinforce this point: technology-related risks include privacy invasion, algorithmic bias, and the “black box” problem. Education-specific risks include student homogenization, the erosion of teacher-student relationships, and deviation from actual learning goals.
These aren’t hypothetical concerns. I see them in my classroom. When a student submits work that sounds polished but reveals no thinking — when the sentences are smooth but the ideas are borrowed — that’s the erosion of learning in real time. The greatest risk of AI in education, as The Conversation recently argued, isn’t cheating. It’s the slow, quiet disappearance of the struggle that makes learning happen.
Bridging the Training Gap
Here’s the uncomfortable truth: teachers aren’t ready. A recent survey shows that while 63% of U.S. teens use AI tools for schoolwork, only 30% of teachers feel confident using those same tools. That gap is a problem — because if we’re not fluent in the technology, we can’t teach students to use it wisely.
The good news is that the professional development landscape is expanding fast. The American Federation of Teachers has partnered with AI developers to train 400,000 teachers. Microsoft’s new Elevate for Educators Credential, built with ISTE+ASCD, aligns to the AI Literacy Framework. Anthropic has partnered with Teach For All to bring AI tools and training to educators in 63 countries.
But credentials and partnerships only go so far. What really changes practice is teachers talking to other teachers — sharing what works, admitting what doesn’t, and building a collective understanding of how AI fits into the messy, human work of teaching.
What Kenny Gets Right
I want to come back to Colleen Kenny’s piece, because she models something important. She transparently disclosed that she used Google Gemini for structure, Claude for style, and Perplexity for fact-checking when writing her article. That’s exactly the kind of honest, intentional AI use we should be teaching — and practicing.
Her core argument is that media literacy has become “education’s most urgent democratic task.” I’d go further: in an international school context, where students straddle multiple information ecosystems and political realities, media literacy isn’t just democratic preparation. It’s survival equipment.
My students will graduate into a world where AI-generated content is everywhere — in their news feeds, their workplaces, their governments’ communications. The question isn’t whether they’ll encounter misinformation. It’s whether they’ll have the critical thinking skills to recognize it.
Where We Go From Here
If you’re a teacher reading this, here’s what I’d say: start where you are. You don’t need to overhaul your entire curriculum. But you do need to be intentional about three things.
First, use AI yourself — not just for fun, but as a professional tool. Use it to draft feedback, brainstorm lesson ideas, analyze student work patterns. Get fluent enough that you can talk about it honestly with your students.
Second, teach source evaluation like your democracy depends on it — because it does. Every discipline has its own version of this. In history, it’s sourcing and corroboration. In science, it’s peer review and replication. In literature, it’s close reading and authorial intent. Whatever your subject, the underlying skill is the same: don’t trust the surface. Go deeper.
Third, be transparent about your own AI use. When I tell my students that I used AI to help draft their feedback, and then show them how I revised it based on what I know about their work, I’m modeling the kind of human-AI collaboration that will define their professional lives.
We’re not going back to a world without AI. But we can build classrooms where students learn to think clearly in a world full of noise. That’s always been the job. It just matters more now than ever.
David Jacobson teaches AP World History at Shanghai American School. He writes about AI, education, and the messy intersection of the two at shouldiuse.ai. Follow him on LinkedIn and Substack.
If you’re thinking through questions like this one — what media literacy looks like when generated content saturates everything — my book goes deeper. The AI Doesn’t Know Your Students is available on Amazon and at shouldiuse.ai/book.
David Jacobson is a high school history teacher. He writes about AI, education, and the messy intersection of the two at shouldiuse.ai.
