Kubrio Logo
Our Newsletter
Kubrio Logo

Rethinking Learning in the Age of AI

Subscribe to our Newsletter

Rethink newsletter

Kubrio World

  • Products
  • Newsletter
  • Open Days
  • Courses

© 2025 Kubrio™. All rights reserved.

Next-Gen Parenting

Your kid used ChatGPT for their essay. Before you panic, ask a different question

Vlad Stan
Vlad Stan
Founder & CEO @Kubrio
Your kid used ChatGPT for their essay. Before you panic, ask a different question
Oct 29, 2025
7 likes
Share:

Was this newsletter forwarded to you? Sign up to get it in your inbox

Not "should they use AI?", but "do they know which mode they're in?" Learning, collaboration, or delegation. The difference is everything.

Your 12-year-old has a history essay due on the Roman Empire. You walk past their room and glance at their screen. ChatGPT is open.

Your first thought: Are they cheating?

Your second thought: Should I stop them?

Here's the thing: you're asking the wrong question.

It's not "should they use AI?" It's: "Do they know which mode they're in?"

Because if your kid asked ChatGPT "Write me a 500-word essay on the fall of Rome" and submitted it, that's one thing.

But if your kid asked ChatGPT "What were the main causes of Rome's fall?", read the explanation, disagreed with some points, asked follow-ups, drafted their own essay, then asked "Does my argument make sense? Where are the weak spots?"—that's something completely different.

Same tool. Completely different outcome.

And here's what most parents (and most kids) don't realize: we're all figuring this out in real time.

You're not alone. OpenAI just released data showing that 10.2% of ChatGPT usage is tutoring and teaching, 8.5% is how-to advice, and 21.3% is seeking information. Adults and kids. We're all students again.

But the biggest category? Practical Guidance: 28.3% of all usage. People asking "how do I handle this?" or "what should I do about that?"

Add it all up and you're looking at nearly 68% of ChatGPT usage being people trying to figure something out.

Here's what the data doesn't show: we have incredible capability, but zero literacy about when to use it.

Because "Practical Guidance" could mean anything. "Help me with this problem" could mean "teach me how to think about this" (Learning), "let's figure this out together" (Collaboration), or "just solve it for me" (Delegation). Same request. Three totally different modes. And most people don't realize they have a choice.


The Real Tension

The problem isn't that people are using AI to learn. The problem is that most of us, kids and adults, don't actually know what we're doing when we open that chat window.

Are we trying to:

  • Learn something we don't understand?
  • Collaborate with AI as a thinking partner?
  • Delegate a task we already know how to do?

All three are valid. But they require completely different approaches. And most people don't know which mode they're in, or which mode they should be in.


Three Modes of AI Literacy

Subscribe to our newsletter

Get the latest insights on AI and human potential delivered to your inbox.

Subscribe Now
Vlad Stan

Vlad Stan

Founder & CEO @Kubrio

View all articles by Vlad Stan →

Let me break this down:

1. Learning Mode: "Teach me this"

You're building foundational understanding. You need the concept explained, broken down, made clear.

Example: Your kid asks ChatGPT to explain photosynthesis because they genuinely don't understand how plants make energy.

Appropriate when: You're entering new territory. You need the mental model first.

The risk: You stop at the explanation and never apply it. The chat ends. No retention. No understanding.


2. Collaboration Mode: "Be my thinking partner"

You have some understanding, but you want to explore, test ideas, get challenged, iterate.

Example: Your kid understands photosynthesis and asks, "Could a plant survive on Mars? What would need to change?" Then they go back and forth with AI, testing assumptions, refining the design, thinking through constraints.

Appropriate when: You're building capability. You're in the messy middle of figuring something out.

The value: You're not outsourcing your thinking, you're sharpening it.


3. Delegation Mode: "Just do this for me"

You understand the domain. You've done it before. You just need execution.

Example: You ask ChatGPT to write a polite email declining a meeting. You know how to write the email, you just don't want to spend 10 minutes on it.

Appropriate when: You understand the system well enough to evaluate the output and you're optimizing for speed, not learning.

The risk: You delegate something you don't actually understand yet. You skip the learning. You outsource your cognitive development.


The Parent's Dilemma

Your kids are using ChatGPT for homework. The question isn't "should they?" The question is: "Do they know which mode they're in?"

Here's the difference:

Scenario A:

Your 12-year-old has a history essay due on the Roman Empire. They ask ChatGPT: "Write me a 500-word essay on the fall of Rome."

ChatGPT writes it. They submit it. Done.

What just happened? They used Delegation Mode for something they don't understand yet. They got a grade, but they didn't learn. And worse, they now think they understand the fall of Rome when they don't.

Scenario B:

Same kid, same essay. They ask ChatGPT: "What were the main causes of Rome's fall?" They read the explanation. Then they ask: "Which cause do you think was most important and why?" They disagree with the AI. They ask follow-up questions. They draft their own essay. They ask ChatGPT: "Does my argument make sense? Where are the weak spots?"

What just happened? They used Learning Mode, then Collaboration Mode. They built understanding. The AI was a sparring partner, not a ghost writer.

Same tool. Completely different outcome.


Why Development is Taking Longer Than We'd Like

This is why we're being so deliberate at Kubrio.

Everyone's racing to build "AI tutors" because the data shows demand. But here's what we keep asking ourselves: are we building a thinking partner, or are we building a delegation machine?

Because the difference is in the details.

If we build a tool that gives kids the answer, we've built a dependency engine. They'll use it like Scenario A. Fast, frictionless, zero learning.

If we build a tool that gives kids a challenge, asks them to struggle, makes them iterate, forces them to think, we've built something different. Collaboration Mode. Real learning.

We want to make sure that when kids use Kubrio, they're not asking "what's the answer?" They're asking "what if I try this?" and "how can I make this better?"

That's hard to design. It takes time. But it's the right thing to build.


The Second-Order Effects

Here's what gets me most excited about this moment:

If nearly 68% of ChatGPT usage is people trying to figure something out, and if we can help people develop the literacy to use AI well, not just frequently, but wisely, then we're about to see learning accelerate in ways we've never seen before.

I'm talking orders of magnitude faster.

Everyone will be better informed. The parent who never understood their kid's math homework can now learn it alongside them. The small business owner who never had time for marketing strategy can ask AI to teach them, collaborate on ideas, iterate in real-time.

Learning used to be bottlenecked by access, teachers, books, or time. Now it's bottlenecked by literacy, do you know how to learn with AI?

The people who figure that out, kids, adults, anyone, are going to run circles around the people who treat AI like a magic answer machine.

I'll write more about this soon. But the TLDR: we're about to see the biggest acceleration in human capability we've ever experienced. If we get the literacy right.


What This Means for You

If you're a parent, here's the shift:

Don't ask: "Is my kid using AI?"

Ask: "Does my kid know which mode they're in when they use AI?"

If you're a builder or founder, here's the shift:

Don't ask: "How can I build a better AI tutor?"

Ask: "Am I building something that teaches people to think, or something that does their thinking for them?"

And if you're anyone else reading this, here's the shift:

You're a student again. We all are.

The question is: are we learning, collaborating, or just delegating?


The capability is here. The tools are incredible. The demand is real.

Now we need the literacy to match.

That's the work.

Vlad

Source: OpenAI Research Paper on ChatGPT Usage Patterns