Last week, one of my AP World History students told me — with complete sincerity — that she’d fact-checked a primary source using ChatGPT before class. Not to cheat. Not to get around the assignment. She wanted to make sure the document was “real” before she spent time analyzing it.
I stood there for a second, not sure whether to be impressed or terrified. And honestly? I think the correct answer is both.
We’ve reached a strange inflection point in education. Students are using AI tools with more fluency than most of their teachers, but that fluency doesn’t automatically translate into literacy. Knowing how to use a thing is not the same as understanding what it’s doing — or what it’s doing to you. And that gap, between fluency and literacy, is quickly becoming the most important challenge in education today.
The Media Literacy Crisis Is Already Here
Colleen Kenny put it bluntly in a recent LinkedIn article: we’re living through “content chaos.” As she wrote, AI and algorithmic curation have created a collapse of shared interpretive frameworks — the common ground we used to rely on for democratic functioning. U.S. media trust has plummeted to an all-time low of 28%, according to Gallup, and Kenny points out that this decline isn’t unique to America. It’s happening “across democracies globally, from Europe to Latin America to Asia.”
Here’s the part that hit me hardest as a teacher: Kenny argues that the average person “can no longer tell the difference between real video and AI-generated video.” If that’s true for adults, what does it mean for our students? And what does it mean for a history class, where the entire point is learning to interrogate sources?
Kenny’s solution is one I deeply agree with: comprehensive media literacy curricula that combine cognitive, emotional, and embodied literacies. And she’s transparent about her own AI use — she disclosed using Google Gemini for structure, Claude for style, and Perplexity for fact-checking — while maintaining human judgment throughout. That kind of transparency is exactly what we should be modeling for students.
What’s Happening in China Right Now
I teach at an international school in Shanghai, and I can tell you: the AI conversation here is moving faster than in most places I’ve seen.
China has made AI a state-mandated part of the curriculum. As NPR reported in early 2026, every student in elementary and middle school in Beijing — and several other districts — is now learning about AI. The curriculum is structured by grade level: third graders learn the basics, fourth graders focus on data and coding, and by fifth grade, students are learning about “intelligent agents” and algorithms. The national goal is to integrate AI into all primary and secondary schools by 2030, and into textbooks, exams, and classrooms at all levels by 2035.
That’s not a pilot program. That’s a civilizational commitment.
Meanwhile, international schools in China are navigating their own path. A recent AmCham China report highlighted how schools across the country are adapting. At Wellington College Tianjin, Dr. Yang notes that AI “boosts efficiency and — most importantly — student outcomes” through personalized content and real-time feedback. They’ve developed their own GenAI policy emphasizing academic integrity and built a Digital Citizenship curriculum that starts in the Early Years.
At Western Academy of Beijing, Stephen Taylor emphasizes that AI should “enhance — never replace — critical thinking, interaction, and overall learning quality.” They’re using tools like Flint for language tutoring and MagicSchool for differentiated materials, but with student workshops on responsible AI use built into the approach.
And then there’s the cautionary voice: Robert Christensen at Optics Valley International School warns about “metacognitive laziness” — the risk that students offload their actual thinking processes to AI. Mahesh Selvaraj at ISA Wuhan echoes this concern, stressing “data-privacy risks, algorithmic bias” and the need to keep “the ‘human touch’ at the center.”
These aren’t abstract debates. These are real educators, in real schools, making real decisions every day about how to use these tools responsibly.
Students Are Worried Too — And We Should Listen
Here’s something that doesn’t get enough attention: students themselves are sounding the alarm. A RAND Corporation survey found that 68% of middle school students now express concern that AI use is eroding their critical thinking — up from 48% just a year earlier. Among high schoolers, 65% share that worry.
But here’s the paradox: even as concerns rise, usage is accelerating. Middle school AI homework use jumped from 30% to 46% between May and December 2025. High school usage rose from 49% to 60%. As Heather Schwartz from RAND puts it, “AI might be giving you a really beautiful explanation about what you can do and how to go about it. It’s still removing that step for you.”
That step — the struggle, the friction, the moment where you don’t yet know the answer — is where learning actually happens. And I see this in my own classroom. When a student comes to me with an AI-generated analysis of a historical document, the writing is often polished, but the thinking is shallow. They’ve skipped the part where they wrestle with ambiguity, where they sit with discomfort, where they build the muscle of interpretation.
What I’m Doing About It (And What You Can Try)
In my AP World History classes, I’ve started treating AI as a source to be interrogated, not a tool to be trusted. When we do document-based questions, I’ll sometimes give students an AI-generated analysis alongside the primary source and ask: Where does this analysis fall short? What context is missing? What assumptions is the AI making?
It turns out this is an incredibly effective way to teach historical thinking. Students have to know the content well enough to catch what the AI gets wrong. They have to understand perspective, bias, and context — not in the abstract, but applied to something that looks authoritative and confident. Because that’s the thing about AI-generated text: it always sounds sure of itself.
I also use AI for my own lesson planning and feedback generation — and I’m open with students about that. When I use Claude to help draft feedback on their essays, I tell them. When I use it to brainstorm activity ideas, I mention it. Transparency isn’t just a nice-to-have; it’s the foundation of any honest conversation about AI in the classroom.
The Bigger Picture: Across Asia and Beyond
This isn’t just a China story. Across Asia, the conversation is accelerating. The Strategic Educational Alliance of South East Asia (SEA2) held its 2026 conference under the theme “AI: Empowering Education,” bringing together schools from Brunei, Malaysia, Indonesia, Singapore, Thailand, the Philippines, and Vietnam. EDUtech Asia 2026 is organized around “Human-centred education, powered by AI and tech.”
But the challenges are real. As educators in Singapore have pointed out, the greatest challenge isn’t acquiring hardware — it’s building genuine capability. Schools need reliable infrastructure, clear policies around safety and ethical use, and teacher training that focuses on societal and ethical issues, not just technical skills. There’s also the localization problem: most generative AI models are trained primarily on Western data, which means they can miss or misrepresent local contexts, histories, and perspectives.
For those of us teaching in international schools, this matters enormously. When an AI tool gives a student a confident but culturally narrow explanation of, say, the Silk Roads or the Opium Wars, that’s not just an academic error — it’s a missed opportunity for the kind of cross-cultural understanding that international education is supposed to foster.
What This Means for Teachers
If you’re an educator reading this, here’s what I want you to take away:
AI is not going away. The question is not whether our students will use it — they already are, at rates that would probably surprise you. The question is whether we’re going to help them use it thoughtfully, critically, and ethically.
That means teaching media literacy not as a unit, but as a habit of mind. It means being transparent about our own AI use. It means creating assignments where AI becomes something to analyze, not just something to use. And it means listening to our students, who are — perhaps more than we realize — already aware that something important is at stake.
My student who fact-checked a primary source with ChatGPT? She wasn’t being lazy. She was trying to be rigorous in the only way she knew how. Our job is to give her better tools for that rigor — and to make sure she knows why the old tools still matter too.
David Jacobson is an AP World History teacher at Shanghai American School. He writes about AI, education, and the future of teaching at shouldiuse.ai. You can also find him on Substack.
Sources cited in this article:
- Colleen Kenny, “Content Chaos: Why AI Makes Media Literacy Education’s Most Urgent Democratic Task,” LinkedIn, February 25, 2026.
- NPR, “In China, AI Is No Longer Optional for Some Kids. It’s Part of the Curriculum,” January 27, 2026.
- AmCham China, “Learning in the Age of AI: How China’s International Schools Are Adapting.”
- Education Week, “Students Are Worried That AI Will Hurt Their Critical Thinking Skills,” March 2026.
- RAND Corporation survey on student AI usage and critical thinking concerns, 2025–2026.
- DigiconAsia, “Is Educational Technology in Asian Primary Schools Evolving with the AI Impact?” March 26, 2026.
