Children growing up with AI still need foundational learning

Wonder Bricks is designing games where children read, calculate, observe, explain, verify, and create with AI instead of outsourcing their thinking to it.

2026-05-25

Children are growing up with AI as part of the default interface. They will ask AI for help, talk to AI characters, and expect software to respond to intent instead of only to buttons. That shift is real. But it does not remove the older educational question. Can a child read a clue, compare quantities, notice a pattern, explain a result, and revise an idea?

At SunnyLabs, we have been studying how AI literacy, foundational learning, game-based creation, and child-facing AI safety should meet inside Wonder Bricks. The answer we keep returning to is simple: Education for children growing up with AI should not mean letting AI think for children. It should mean giving children better loops for thinking with AI.

The next generation does not need games that hide learning under worksheets. It needs games where observation, reading, calculation, explanation, and revision are part of play.

Foundational skills are still the floor

The World Bank defines foundational learning as basic literacy, numeracy, and socio-emotional skills, and describes these skills as the basis for lifelong learning, later school success, work, and citizenship.¹ That matters for AI-era products because AI can make weak foundations less visible. A child may produce a fluent answer with help from a model while still not understanding the text, the quantity, or the cause-and-effect relationship underneath it.

So Wonder Bricks does not treat "educational game" as a thin layer of quiz questions. In our game planning, every learning-oriented world should have a clear academic core: reading clues, comparing numbers, testing a science idea, building with spatial reasoning, debugging a rule, or explaining what changed after a choice.

AI literacy belongs inside the play loop

UNESCO's student AI competency framework frames students as responsible AI users and co-creators, with competencies across a human-centred mindset, AI ethics, AI techniques and applications, and AI system design.² AI4K12 makes the same idea concrete for younger learners through five big ideas: perception, representation and reasoning, learning, natural interaction, and societal impact.⁵

For Wonder Bricks, that does not mean children need a lecture about machine learning before they can play. It means the game itself should make AI understandable. An AI friend may misread a clue. A Kiki revision may need a clearer instruction. A creature may "learn" from repeated examples. A child may compare an AI suggestion with what actually happened in the world.

Those moments teach an important habit: AI output is something to inspect, test, and improve, not something to obey automatically.

The game loop we want: observe, predict, test, explain, revise

OECD's PISA work on creative thinking focuses on how students generate, evaluate, and improve ideas.⁴ That maps naturally to games. A child can notice that one bridge keeps falling, predict that a wider base will help, test the build, explain the result to a Wonder Friend, then ask Kiki to modify the design.

That is the educational loop we want more of in Wonder Bricks: observe, predict, test, explain, revise. It is more durable than a single correct answer, and it fits the way children who grow up with AI already expect software to behave. They try something, talk about it, and change it.

Wonder Friends should be thinking partners, not answer machines

AI tutoring research suggests that intelligent tutoring systems can support learning, but the design details matter: subject domain, feedback style, learner control, and the surrounding classroom or family context all shape outcomes.⁷ The U.S. Department of Education also emphasizes keeping humans in the loop as AI enters teaching and learning.⁶

That is why Wonder Friends are more promising as bounded thinking partners than as instant answer machines. A good Wonder Friend asks, "What did you notice?" It gives a hint before a solution. It asks a child to compare two strategies. It helps turn a messy explanation into a clearer one. The child still does the important work: reading, counting, observing, choosing, explaining, and trying again.

Safety is part of the educational design

UNESCO's guidance on generative AI in education calls for a human-centred, age-appropriate approach, including attention to privacy, ethical validation, and pedagogical design.³ Common Sense Media's 2025 research on AI companions shows why this matters: teen use of AI companions is already widespread, and the risks include privacy sharing, emotional dependence, uncomfortable interactions, and blurred relationship boundaries.⁸

For Wonder Bricks, education and safety cannot be separate roadmaps. If an AI character helps a child learn, it must also keep a clear AI identity, avoid secretive or dependent relationships, respect memory and deletion controls, and stay inside age-appropriate boundaries. Trust is not a wrapper around the learning experience. It is part of the learning experience.

What this means for Wonder Bricks content

Wonder Bricks is developing content around a practical principle: every game for children growing up with AI should give children a reason to think. A racing game can teach distance, time, and strategy. A building game can teach geometry and cause-and-effect. A story world can teach reading, vocabulary, and narrative consequence. A puzzle can teach rules, conditions, and debugging. A science world can teach variables and observation.

The AI layer makes these worlds more alive, but the learning value comes from the loop. Children should make a guess, see a result, talk about it, and improve the world. That is where Wonder Bricks can be different: not just more content generated by AI, but more children learning how to create, question, verify, and revise with AI.