AI Can Give Children Answers. But Does It Build Thinking Skills?

"Help me choose a theme for my summer research project." "Tell me the answer to my homework." "Write my book report." Children are starting to ask AI the kinds of questions they once typed into a search box.

When a child speaks to a PC or tablet and says "do this," text, pictures, and even small programs come back in seconds. Watching it, you might feel that AI has become a tool anyone can use, one that requires no special skill at all.

But "being able to use it" and "being able to use it well" are not the same thing. Even with the same tool in front of them, people get very different results depending on what they ask for and how they revise what comes back.

This is not only a question about children's learning. In adult workplaces, too, there is already a gap in how well people can use AI. In an analysis published by OpenAI in 2025, the top 5% of AI users — those who had adopted it most heavily — exchanged about six times as many messages with AI as the median employee at the same workplace. Even when the same tool is handed out, some people fold it into their daily work, while others barely touch it.

There is also a gap in understanding. In a literacy assessment run in spring 2025 by the Generative AI Utilization Promotion Association (GUGA) in Japan, 50% of those holding the "Generative AI Passport" certification scored in the upper half of the distribution, compared with 14% of the uncertified. Nomura Research Institute's 2025 survey likewise found that while generative-AI adoption among companies reached 57.7%, the single largest challenge cited — by 70.3% of firms — was "a lack of literacy and skills." The tool is spreading, but the ability to use it in work and learning is not spreading at the same pace.

The Divide Isn't About "Operating" the Tool

Typing on a keyboard and speaking to an AI is something anyone can do right away. But using AI well is not mainly about knowing which buttons to press. It is about what you think through before giving an instruction, and how you judge the answer after it comes back.

The first is the ability to put your goal into words. Unless you can define for yourself "what I want to make" and "what kind of answer I'm after," you cannot give AI a precise instruction. The second is the ability to reframe the question. When the first answer falls short, you work out what to change and how to get closer, then rebuild the instruction. The third is the ability to doubt what comes out and verify it.

Generative AI sometimes returns information that sounds plausible but has no basis in fact. This is often called a "hallucination." Without the habit of checking the output against other sources, a useful tool can also become a source of repeated mistakes.

Which Is Why Not Everyone Uses It the Same Way

Putting a goal into words, rebuilding the question, checking the output: these are the parts we cannot simply hand over to AI. They also do not appear automatically just because a tool has been distributed.

The point is not that relying on AI is bad in itself. The risk is building a habit of taking the answer before doing the thinking. With homework, for example, if a child gets the answer from the start, there is less time spent reading the question, sorting out the conditions, trying something, getting it wrong, and thinking again. AI can help children build their thinking skills, but used carelessly, it can also take away the very chances they need to develop those skills.

That is why the spread of AI can widen the gap between those who use it well and those who do not. Two people can hold the same smartphone, with one using it for research and organizing information while the other mostly watches videos. Having the tool and expanding what you can do with it need to be treated as separate things.

A Capacity You Can Build From Childhood

These capacities are not acquired all at once in adulthood. From childhood, what matters is not producing an answer quickly but putting into words what you actually want to do; not swallowing the answer whole but checking "is this really true?"; and, when things do not work, reframing the question itself. Those experiences become a foundation over time. This is not only a matter for elementary schoolers — for middle and high school students, too, it is something they can begin building now.

At Digital Kodomo BASE, we want children to use AI and PCs for their own purposes, rather than be led by the tools. We want them to grow into adults who can think, choose, and keep learning as they make their way through life.

That is why we are still working through how to help children build logical thinking skills and the ability to reframe their own questions. We value the experience of deciding, trying, and checking for themselves, and we will keep building those experiences with the children who come to us.