A friend in Cairo sent me a voice note last week: "My son asked Gemini how planes fly, got a perfect answer in six seconds, and moved on. Is that learning?"
Not yet. But it can be. The risk of AI for curious kids isn't laziness — it's premature closure: the habit of treating the first answer as the final one. Here's what works.
1. Make "why" a two-step game
After any AI answer, the rule in our house is: one more why. Not to interrogate the model — to interrogate the kid. "The AI says planes fly because of lift — why does tilting the wing make lift?" The second question is where real understanding lives.
2. Offer pre-AI guesses
Before you open an app, ask your child to guess. Write it on a Post-it. Then check. Kids who guess first remember 3x more of what they read (Rosenbaum et al, 2019). It's also where the pride lives.
3. Celebrate wrong guesses, loudly
Every wrong guess is a proof the child is thinking, not googling. Make a micro-ritual out of it: "Excellent wrong guess! Let's find out why." It's disarming, and it builds a kid who is unafraid to think out loud.
The MENA-specific angle
In many MENA households, being "right" carries extra weight — especially in mixed-age family gatherings. Start in private. Model wrong guesses yourself. Once the home dynamic is safe, the curiosity moves with them everywhere — including into Arabic class and into the questions they ask at جدّتي's house.