How Do We Develop AI Fluency in Children?

Young Children Need Guidance

Independence develops gradually. So does judgement.

We don’t expect a three-year-old to read a book independently before we introduce books to them.

A teacher reads to the child. The child learns how a book works — that the pages turn, that words carry meaning, that pictures tell part of a story, and that stories have beginnings and endings. Over time, the child begins to recognise words, understand how reading works and eventually learns to read independently.

The same is true of writing. We don’t wait until a child can write before giving them paper and a pencil. We let them draw, make marks, experiment and gradually develop the skills needed for writing.

AI should be approached in much the same way.

There are books that are inappropriate or harmful for young children. We don’t therefore ban books. We rely on teachers and parents to select age-appropriate books and guide children in how they experience them.

AI is no different.

At preschool age, children are still developing the wisdom, judgement and self-control needed to make important decisions. We therefore do not expect young children to use AI independently or to decide for themselves whether an AI response or recommendation is appropriate.

Teachers must supervise children’s use of AI and remain responsible for deciding when and how AI is used, what children are exposed to, and how they are guided to understand what AI produces.

But supervision does not mean keeping children away from AI. It means introducing AI appropriately.

The child is therefore learning about AI through using AI, rather than simply being told about it or forbidden from using it.

If AI produces something inaccurate, unexpected or simply wrong, the teacher can use that moment to explain: “AI can make mistakes too.”

What Should Children Gradually Learn About AI?

AI fluency is not about mastering today’s AI tools. The tools will change.

It is about gradually understanding what AI can do, what it cannot do, how people can work with it, and where human judgement remains essential.

1. AI as a Tool — To Augment What Children Can Do

AI can give children capabilities that would otherwise be difficult — or even out of reach — for them.

For example, a child can create a complete storybook with their own story, characters and ideas, and use AI to help create the illustrations. Before AI, producing a complete storybook with original illustrations would have been much more difficult for a young child.

The same idea can be seen in our Mother’s Day projects. A child can describe what makes their mum special — her personality, the things she likes to do, or what the child loves about her — and AI can help turn those ideas into a unique song created especially for her.

The child provides the ideas and makes the creative decisions. AI gives the child new ways to bring those ideas to life.

AI is not replacing the child’s creativity; it is making forms of expression that were previously difficult or inaccessible much more achievable.

2. AI as a Revision Tutor

AI can personalise practice in ways that were difficult to provide at scale before AI.

Our AI Chinese programme uses AI to remember what each child has learned and identify the words that individual child still needs to practise. Instead of every child simply receiving the same revision, the system can provide practice that is more closely matched to the child’s own learning needs.

Our AI Maths programme applies a similar principle to mathematics, providing children with additional practice based on their individual learning.

Before AI, providing this degree of personalised practice for every child would have required a teacher to track each child’s learning and prepare different practice continually. AI makes it possible to provide this additional individualised practice much more efficiently.

AI therefore does something that was previously difficult to do at scale: it can provide each child with more personalised opportunities to practise.

But AI does not replace the teacher. The teacher remains responsible for teaching, understanding the child and deciding what support the child needs.

3. AI as a Tutee

Tutee means student, learner.

AI and robots can also become the learner.

When children programme a robot, they give the machine instructions, observe what happens and correct their instructions when the robot does not behave as intended.

This creates an unusual learning experience for young children. Instead of always being the one who is told what to do, they get to tell something else what to do.

And the robot will not laugh at them when they make a mistake.

If the robot does something unexpected, the child can try again, change the instructions and see what happens.

The child becomes the teacher.

This helps children learn to communicate clearly, test their ideas, learn from mistakes and persist until they get the result they want.

4. AI Can Make Mistakes

AI can be remarkably capable — but it can also be wrong.

Children can begin to develop the habit of asking: “Is this right?”

But at preschool age, they should not be expected to verify important information independently. Teachers help children check, question and understand what AI produces.

When something seems wrong or uncertain, children should learn to stop and seek help from a trusted adult.

They learn that AI can make mistakes too.

5. AI Can Recommend — Humans Must Decide

AI can increasingly make recommendations about what we should read, watch, learn or do.

But a recommendation is not a decision. Important decisions remain with the adults responsible for the child.

As children grow, they can gradually learn to question recommendations, consider alternatives and understand why a particular decision is appropriate.

AI Fluency Is About the Relationship Between Child and AI

We want children to understand that AI can be powerful without being infallible; useful without being in control; and intelligent without replacing human judgement.

AI can provide capabilities and possibilities. The child remains responsible for what they create, believe and decide — within an environment where the teacher provides the guidance and boundaries appropriate for their age.

The more capable AI becomes, the more important it is for children to develop both the ability to work with it and the Human Intelligence to know when and how it should be used.

How Do We Teach AI?

Knowing what capabilities we want children to develop is only half the story.

At ChildFirst, we develop AI fluency through carefully designed, age-appropriate experiences in coding, robotics, creative projects and personalised learning — with the teacher guiding the experience and the child actively participating.

How Do We Teach AI →

Explore how ChildFirst puts these principles into practice.