What a youth camp in Garam Chashma revealed about the real AI divide, and it isn’t the one we keep talking about.
I spent twenty days this summer running youth camps in Garam Chashma, Chitral, with students aged fourteen to twenty. I went in expecting to talk a lot about access bandwidth, devices, career goals, challenges and the usual checklist for a remote valley like this one.
A girl finished a camp assignment task on her phone in about 2 minutes. She had asked an AI tool for the answer, copied it, and was ready to move on. I asked her a simple question: what did she think of the answer it gave her?
She looked at me, genuinely puzzled. “Sir, theek tha” (Sir, it was fine), she said, and waited for the next question, not because she couldn’t think harder about it, but because no one had ever asked her to.
The Assumption I Walked in With
I had assumed, like most people writing about AI in remote Pakistan do, that the gap would be about connectivity and awareness, whether these students had signal, whether they’d even heard of the tools urban teenagers use daily.
The camp corrected that assumption within the first two days. Nearly every student had a smartphone. Most used AI tools for homework, for content, for shortcuts through boring assignments, as naturally as their peers in Karachi or Lahore do. The access gap I had in my mind barely existed at the level I expected.
What did exist, and what no checklist had prepared me to notice, was something quieter: not one student I spoke with had ever been asked what they made of an AI-generated answer. Not by a teacher. Not by a parent. Not by anyone.
The Real Divide: Guidance, Not Access
There is a growing body of research on this exact distinction, and it matters here more than any statistic on internet penetration. A recent systematic review of how ChatGPT affects critical thinking in higher education found that outcomes depend less on whether students use AI and more on how that use is structured, whether a rubric, a prompt, or a mentor pushes them to compare, justify, or revise what the tool hands back. Left unstructured, especially for less experienced users, AI interaction tends to slide into what researchers call cognitive offloading, the model does the thinking, the student does the copying.
The tool itself is not the variable. The presence of an adult who asks the next question is.
That is the piece nobody in Garam Chashma had. Not from a shortage of intelligence or curiosity in the students, I met plenty of both, but from a structural absence of anyone whose role it was to sit beside a teenager and ask what she thought of what the machine just told her.
Why Garam Chashma Specifically
Urban Pakistani households are not solving this problem well either. But they have partial scaffolding that remote valleys like Chitral simply don’t. A tuition centre that at least occasionally pushes a student to explain their work. An English-medium school with a teacher trained to ask a follow-up question. A career counsellor, even an informal one, who has seen enough students to know when an answer sounds copied rather than understood.
Reporting on Pakistan’s AI landscape has noted a sharp access disparity between urban and rural students, and separately, that women in particular face far lower daily internet use, both real and worth noting. But in Garam Chashma, the students I worked with mostly had phones, data, and tools. What they didn’t have was a single adult in the room whose job was to model how to question, verify, or build on what AI gave them. Multiply one puzzled “theek tha” by a generation of remote-area youth with connectivity and no equivalent of a tuition centre, and the shape of the real problem becomes clear.
This is territory AKU-IED’s Professional Development Centre in Chitral already understands well, not just infrastructure, but the mentorship and professional development scaffolding that turn access into actual capability. It is exactly the kind of gap that community-based capacity-building, rather than another connectivity drive, is built to close.
The Question Worth Asking
For years, the instinct in conversations about remote Pakistan has been to ask how we get more bandwidth to more villages. That instinct isn’t wrong, but it has quietly become the whole conversation, and it is answering a problem this camp mostly didn’t have.
The better question, the one I keep returning to since Garam Chashma, is smaller and harder to solve with a tower or a SIM card: who is going to be the person that asks a teenager the next question? Who sits with her after the assignment is submitted and asks not whether the answer was right, but what she thought of it?
That person doesn’t need to be an AI expert. She needs to be present, curious, and willing to ask a question that has never been asked before.