Another hotel ballroom. Another plate of cold samosas. Another room full of people in expensive suits trying to figure out how to automate the one thing we still haven’t quite mastered: teaching a child how to think.

The latest gathering of "stakeholders" met this week to talk about the National Education Policy (NEP) 2020 and its inevitable collision with Artificial Intelligence. They’re calling it "alignment." It sounds professional. It sounds planned. It mostly sounds like a bunch of software vendors realized that India’s 250 million students represent the largest untapped data set on the planet.

Let’s be clear. The NEP 2020 is a dense, ambitious document that wants to overhaul how India learns. It wants to move away from rote memorization and toward something more fluid. Enter the AI crowd, smelling blood in the water and venture capital in the air. Their pitch is simple: why pay for better teacher-to-student ratios when you can just give every kid a chatbot?

They’re calling it "personalized learning paths." In plain English, that’s an algorithm deciding if an eight-year-old in a rural village is "vocational material" before they’ve even hit puberty. It’s a tracking system with a better UI.

There’s a specific kind of friction here that the brochures don't mention. Take the "National Educational Technology Forum" (NETF). It’s supposed to be the bridge between policy and code. But during these discussions, the friction wasn't about pedagogy. It was about the bill. Integrating high-end LLMs into the public school system isn't cheap. We’re talking about a multi-billion-dollar pivot toward proprietary tech. If the government bakes a specific company’s API into the national curriculum, they aren't just buying a tool. They’re handing over the keys to the country’s intellectual infrastructure.

The EdTech CEOs in the room were grinning. Why wouldn’t they be? The NEP 2020 encourages "coding from a young age" and "AI integration." To a guy with a SaaS startup, that looks like a state-mandated customer base. They don't talk about the digital divide anymore. That’s old hat. Now they talk about "equitable access to intelligence." It’s the same problem with a new hat.

I spoke to a middle-school teacher from a government school outside Noida. She wasn't at the meeting. She doesn't have time for ballrooms. She has 55 students in a room designed for 30 and a ceiling that drips when it rains. I asked her what she thought about "AI alignment" with the policy.

"I'd settle for a printer that has ink," she told me.

That’s the reality the stakeholders ignore. You can’t run a "hyper-personalized AI tutor" on a tablet that hasn't been charged because the school’s power is out. You can’t "align" with the future when the present is still waiting for basic plumbing.

There’s also the matter of the data. The NEP 2020 talks about a "holistic progress card." It’s a digital record that follows a student from preschool to university. Now, imagine plugging an AI into that. Every mistake, every late assignment, every failed quiz—all fed into a black box to "optimize" the student’s future. It’s a permanent record on steroids. If we’ve learned anything from the last decade of tech, it’s that "optimization" is usually just a polite word for "pigeonholing."

The discussion ended with a lot of talk about "synergy." They want to create "intelligent content" that adapts to the learner. It sounds nice until you realize that "intelligent content" is just a way to replace a human who cares with a script that doesn’t.

By the time the last PowerPoint slide flickered off, the consensus was reached: the integration is inevitable. The policy demands it. The market craves it. The bureaucrats need the "modernization" checkmark on their performance reviews.

The kids? They’ll be the ones beta-testing this mess. They’ll be the ones navigating a world where their teachers are increasingly sidelined by "smart" dashboards that prioritize throughput over actual understanding.

As the stakeholders filed out of the ballroom, I noticed a pile of printed copies of the NEP 2020 left on a side table. They were pristine. Unread.

We’re moving at light speed to automate an education system that we haven't even finished building yet. We’re obsessed with the "AI" part of the equation and seemingly bored by the "Education" part. But hey, at least the dashboards will look great in the annual report.

Who actually benefits when the algorithm decides what a child is worth?