Thoughts on Questioning in the Age of AI
In an age when AI has all the answers, knowing what to ask next is the only thing it cannot do for you.
My son turned two last year. Since then, his standard response to almost everything is a “Kaa?” That’s “why?” in Marathi. He asks it relentlessly, until he gets a satisfying answer. And sometimes even when the answer starts with “because” and the reason has technically been given, he asks again. It’s like a never ending loop 🤣. It’s occasionally annoying. But we love it, because that is precisely how toddlers learn, and we never want him to stop.
What he does instinctively has a name in business schools: the why ladder. Keep asking why until you reach the root of a problem. It is Socratic questioning in its purest form. No training required. Children are born with this.
But somewhere along the way, most of us stop questioning.
I noticed this about myself too. As I grew up, the number of questions I asked each day dropped significantly. It made me wonder: did I lose my childlike curiosity? Is the quantity of questions as important as their quality? And how important is questioning as a skill in the age of AI? 🤔
For most of human history, the bottleneck was answers. Finding good information was the hard problem. Libraries, pundits, experienced people … all designed to bring answers to the people who needed them.
AI solved that problem in 2022. (ChatGPT was born that year!)
The bottleneck has now moved entirely to the other side of the equation: the question.
But have the schools caught up on it? Exams still test recall. Homework often tests answer-finding. Rubrics still reward the correct response. This is not a failure of intention, it is a system faithfully doing what it was designed to do.
But that system design is now outdated.
When education focuses on answer finding, it neglects the skill AI makes most valuable: knowing what to ask.
Research puts numbers to something most of us sense intuitively. Pre-schoolers ask dozens of questions a day. By middle school, that number drops sharply. At home, children ask questions constantly. In classrooms, the frequency falls dramatically.
This is not because teachers are suppressing curiosity. A teacher cannot run a Socratic dialogue with 30 students at once. The culprit is a system of mass education, not an individual educator.
But naming the structural cause does not make the outcome less serious.
There is a difference between querying and questioning.
A query is: “What is photosynthesis?”
A question is: “Why does photosynthesis rely on green pigment specifically? What would change if plants had evolved a different system?”
The first gets you a summary. The second starts a conversation that might go somewhere.
A Google search is sufficient for a query. Most people interact with AI exclusively in query mode. Imagine using the most powerful thinking partner in human history as a glorified search engine. 😅
This is also where “prompt literacy” only goes so far. Writing better prompts is useful. It helps you get better outputs from AI.
But question literacy is different.
Prompt literacy sharpens the tool. Question literacy sharpens the person.
You can write an excellent prompt and still accept the output uncritically. Question literacy is the habit of asking with intent: to clarify, challenge, probe assumptions, explore alternatives, or decide what to do next.
Like I mentioned what Karpathy did in my previous article [https://nehal.work/notes/the-curious-case-of-critical-thinking/]. That is the part we need to teach!
The why ladder is not just a questioning technique. It is first-principles thinking in practice. Every time a child refuses to let the first answer be the last one, they are challenging the assumption underneath the answer. They are decomposing the problem (helps build computational thinking). They are building understanding from the ground up.
We tested this in a pilot cohort at Lernok. When kids used the why ladder consistently, something interesting happened. They started breaking problems down to their most basic form without being asked to.
The questioning habit became the thinking habit.
You cannot teach first-principles thinking by announcing it as a concept. You build it by protecting the reflex to keep asking.
Some people argue that curiosity cannot be taught. That it either stays alive or it does not, shaped by environment and the adults around a child. Turn it into a formal technique, the argument goes, and you kill the spirit you were trying to protect.
I think that is a false choice.
The why ladder is scaffolding, not a replacement for curiosity. Genuine curiosity, when given structure and permission, compounds rather than diminishes.
My son does not need the technique. He is 3yo. But the 14yo who has been gradually trained out of asking might need the scaffolding to find the reflex in their ‘thinking muscles’ again.
The practical version is simpler than it sounds.
After a child uses AI for anything, ask one follow-up question: “What do you think it got wrong?” or “What does that make you want to ask next?” That’s it. Just the habit of going one question further than the first answer.
My son is going to hit formal school eventually. The system will do what it was designed to do. But right now, every “Kaa?” he asks still gives me hope that he will have the appetite for questions in a world full of answers at fingertips.
— Nehal