In recent years, more and more parents have begun to feel anxious: now that AI has arrived, what will children rely on in the future?

In the past, we often believed that if children had good grades, learned enough knowledge, and solved problems quickly enough, their future would be more secure. But now, many things are changing. When children do not understand something, AI can explain it immediately. When children cannot write something, AI can generate it quickly. Even summarizing key points, outlining ideas, and suggesting methods are becoming things that AI can do increasingly well.

This has led me to rethink what truly matters. In the future, the key may not be how much knowledge children have learned, but whether they can think and judge for themselves within an overwhelming amount of information. Knowledge can be accessed quickly. But how a person understands problems, breaks them down, and makes choices when there is no standard answer will determine how far they can go.

I have come to believe that in the AI era, what parents should spend more time cultivating is not children’s familiarity with answers, but the quality of their thinking.

The quality of thinking is not simply about whether a child is smart. It is about whether, when facing a problem, the child can pause and ask: What is this question really asking? Is there another method I can try? Why does this answer look right but may not be right? If I fail, what can I adjust next?

This connects closely with several ideas in educational psychology.

First, I strongly value the concept of a growth mindset. Some children immediately say “I cannot do it,” “I do not know how,” or “I am just not good at this” when they meet difficulty. But when children gradually develop a growth mindset, they begin to see difficulty as a process rather than a conclusion. They no longer think only “I cannot do this now.” They begin to think, “I cannot do this yet, but I can try again.”

That word “yet” is important. It helps children understand that ability is not fixed, and that thinking skills are not unchangeable from birth. Very often, what children lack is not an answer, but the confidence to keep thinking.

Another idea I care deeply about is metacognition. Simply put, this means children are not only solving a problem, but are also aware of how they are thinking. When working on the same question, some children simply keep calculating. Others pause and ask themselves: Where am I stuck? Did I misunderstand the question? Is my method correct? Is there another way to think about this?

This ability to observe one’s own thinking process is extremely important for the future. AI can provide answers, but it cannot build a child’s awareness of their own thinking state. A person who can reflect on their own thinking is more likely to maintain learning ability and less likely to be led passively by tools.

The third idea that resonates with me is self-determination theory. It reminds us that people usually sustain motivation not because they are constantly pushed, corrected, or demanded, but because they experience three things in the process: I am capable of doing something, I can make choices, and what I am doing is connected to me.

This is very meaningful for parenting. If a child is always arranged, corrected, and required to produce standard answers, they may become very compliant over time, but not necessarily good at thinking. But if children are often encouraged to express their own understanding, explain their reasons, and compare different approaches, they gradually feel that their ideas have value and that they are capable of thinking for themselves. This sense of initiative is one of the abilities that will be difficult to replace in the future.

Now, I am less eager to ask children, “What new knowledge did you learn today?” I pay more attention to whether they tried first when facing a difficult problem; whether they could explain what went wrong after making a mistake; whether they could give reasons when their answer differed from someone else’s; and whether they dared to ask, think, and revise when facing uncertainty.

These things may not look like grades, but they may matter more than grades.

In the future, knowledge will update faster, and tools will become more powerful. What is truly scarce may not be knowing, but thinking. A child who can think does not never make mistakes. They know how to adjust after mistakes. They do not always produce answers quickly. They are willing to break problems down slowly. They do not need to be stronger than AI in everything. But they will not treat AI’s output as the only truth.

We cannot fully predict the world children will face ten or twenty years from now. But we can decide whether today we help them become people who can think independently. If children are only passive receivers, used to others organizing, judging, and concluding for them, then as tools become stronger, they may lose more of their own agency. But if children are encouraged from an early age to observe, ask, analyze, compare, and revise, then no matter how quickly the world changes, they will have a stronger chance of standing firmly on their own.

Rather than rushing children to learn more earlier, I increasingly believe that helping them learn how to think, how to face problems, and how to find direction amid uncertainty is the more lasting gift.

Knowledge will update. Tools will upgrade. Answers will change. But the ability to think actively, analyze problems, and continually revise oneself is a foundational capacity that can accompany a child for life.

Suggested meta description: In the AI era, children need independent thinking, metacognition, growth mindset, and judgment more than simply learning earlier or memorizing more.