It's sometimes hard to question your own assumptions, and everyone has blind spots, but children are often uniquely gifted at spotting every potential flaw in an explanation and cutting to the heart of a situation with a question or two. They are experts at spotting missing pieces, perhaps because they have not yet learned to politely ignore them.
After some recent experiences with precisely that phenomenon, I decided to experiment with applying it to AI prompts. I asked ChatGPT to act like a curious eight-year-old whose job was to find flaws in my ideas.
Specifically, I started a conversation by telling ChatGPT to: "Pretend you are an intelligent, curious child. Look at my idea and ask the kinds of questions a kid might ask. Focus on simple things that don't make sense or seem incomplete."
Boring Saturday
The AI's response was gratifying even as it became much harder to hide from obvious questions. For instance, I occasionally create ambitious weekend plans that look fantastic on paper and exhausting in reality. I shared the schedule with ChatGPT and asked for its childlike review.
The response reminded me of my own thoughts as a kid: "If Saturdays are supposed to be fun, why are you putting all the boring things on Saturday?"
The chatbot continued by asking why some tasks were scheduled for that specific day instead of another day, and why there was no time left for anything spontaneous. After looking at the schedule again, I agreed I'd overpacked it and resolved to clear up some time to relax and not necessarily be productive.
Solving streaming
With that free time and a resolution to relax and watch a movie, the next question was what to watch. After scrolling through menus, considering several options, rejecting them all, I was no closer to a decision than when I started. Again, I asked ChatGPT for the child's perspective.
"If you wanted to watch something, why didn't you watch something?"
That was difficult to argue with. The chatbot asked why I was looking for the perfect movie instead of a movie. Why spend more time deciding than watching? All good things to consider, so I looked at my short list, picked one at random, and enjoyed the film.
Some of the most useful questions are not the complicated ones. The same goes with prompts. Most chatbot interactions involve asking for better answers, deeper analysis, or more expertise. This prompt worked because it reduced everything to first principles.
Of course, no chatbot is genuinely thinking like a child. It is still an AI generating responses based on patterns in its training data. Yet asking it to adopt that perspective creates a useful filter. Children do not necessarily have better answers than adults, but they do sometimes have better questions.
Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds.
