AI Literacy · Elementary School

Teaching children to think about thinking.

Kids Thinking Club is an AI literacy and critical thinking curriculum for elementary school children — designed to build the conceptual tools kids need to understand, question, and engage thoughtfully with the AI-shaped world they are growing up in.

Ages 6–10 Hands-on & discussion-based No screens required Ethics built in from the start Designed for curious kids
Get in touch about your school →
Why this matters

AI literacy is civic education.

Children growing up today will navigate AI-shaped systems their entire lives — in school, at work, in healthcare, in public life. They deserve to understand what these systems are, how they work, and who they are built for.

But AI literacy isn't just a technical skill. It's a thinking skill. Kids Thinking Club starts with the most important question of all: What is intelligence? — and builds from there.

Start with wonder, not toolsKids don't need to use AI to understand it. They need to think carefully about what it can and can't do — and why that matters.
Ethics from day oneBias, fairness, and the limits of pattern-matching aren't advanced topics. They're exactly what young children are ready to grapple with — on the playground, in stories, in their own lives.
Questions over answersThe goal isn't to teach children what to think about AI. It's to give them the vocabulary and confidence to think for themselves.
Belong before you learnEvery child in the room should see themselves as someone who can understand and question powerful systems. That confidence is built deliberately.
How it works

Designed for how children actually learn.

Each unit combines a big conceptual question with a hands-on activity and a structured discussion. Children move between thinking and doing — building intuition before encountering the formal idea.

Big conceptual questions

Each unit opens with a question children already have feelings about — Is a wolf intelligent? Can a computer see? — before introducing the formal concept.

Hands-on activities

Children learn by doing — pixel coloring to understand how computers see, sorting games to grasp pattern recognition, role-playing decision trees to feel how AI systems work.

Structured discussion

Reflection questions are built into every session — not as assessment, but as practice. Children learn to articulate uncertainty, notice gaps, and push back on confident-sounding claims.

Ethics woven throughout

Questions of fairness, missing data, and who gets left out aren't saved for later units. They arise naturally as children start to understand how systems learn — and what they miss.

Sample units

Three units from the opening arc.

The full curriculum spans multiple units across the school year. Below are the first three — click any unit to explore the sessions.

1
What is Intelligence?
Unit 1 · January 2026
▼

The opening unit asks the question that underpins everything: what does it mean to be intelligent? Children explore intelligence across species — wolves, octopuses, bees, fungi — before confronting whether computers can be intelligent too. The unit ends with a working definition children have built themselves, and two “conundrums” that complicate it.

Is grass intelligent?
Opening provocation — exploring the edges of what intelligence means before defining it. Children argue both sides.
How are these animals intelligent?
Wolf, cockroach, octopus, beehive, pig, virus, parrot, shark — how does each show intelligence differently? Children discover that intelligence is not one thing.
What do these things have in common?
Ants, fish schools, beehives, mycelium networks — collective and distributed intelligence. No single brain, but the system is smart.
Building a working definition
Key ideas children arrive at: learning from experience, adapting when conditions change, deciding between options, solving new problems, fitting the environment. Stephen Hawking: “the ability to adapt to change.”
How do computers learn?
A computer is a very fast rule-follower. Humans wrote all the rules. “If the light is red → stop.” But what happens when the rule doesn’t fit? Scientists tried a new idea: show the computer 10,000 pictures of cats instead of writing every rule. Patterns start to appear.
Can we teach a computer to make hummus?
A concrete activity: can you write every rule a computer would need to make hummus? Children discover just how hard it is to make rules for things humans do intuitively.
What is Artificial Intelligence?
“Artificial intelligence is when people teach computers to learn, instead of telling them every rule.” Introduced through the cat picture example — a computer that has seen 10,000 cats but doesn’t know what a cat is.
◆ Discussion conundrums The Game Conundrum and The Puppet Conundrum — two scenarios designed to stretch the definition children built and expose its limits. Is a chess computer intelligent? Is a puppet that learns your moves?
Picture books as discussion openers
Eye Spy: Wild Ways Animals See the WorldGuillaume Duprat — opens the question of how different creatures perceive and make sense of the world. Perfect for the “how are these animals intelligent?” discussion.
Grandmother Fish: A Child’s First Book of EvolutionJonathan Tweet — introduces the idea that intelligence evolved differently across species. Grounds the animal intelligence discussion in shared ancestry.
Becoming a Good CreatureSy Montgomery — explores what it means to be intelligent and ethical across animal species. Rich for the “what do these animals have in common?” slide.
Zips and Eloo Make HummusLeila Boukarim — used directly in the “can we teach a computer to make hummus?” activity. Children discover just how hard it is to write rules for something as intuitive as making food.
2
How Does a Computer See?
Unit 2 · January 2026
▼

Unit 2 digs into perception — how do humans sense the world, and how is that different from how computers do it? Children discover that computers don’t see objects at all: they see numbers. The central activity is pixel coloring: children color grids by number at three resolutions to experience directly what it means to “see” through a grid of values.

Revisiting the Game Conundrum
Opening callback to Unit 1’s unresolved question — building continuity across sessions.
Sensing the world
Human senses as a model. Activity: explore three objects using only one sense at a time. Which sense felt hardest alone? What clues helped most?
What does the world really look like?
Bridging from human perception to the question of computer perception. How do you think computers sense the world?
What’s in a “cat”?
What do humans actually see — and understand — when they look at a cat? Softness, mood, memory, meaning. A computer has none of this.
Selective attention: how many passes?
The famous awareness test — how many times did the white team pass the ball? Children discover that even humans miss things when focused. Computers miss things differently.
How does a computer see?
Computers don’t see objects. They see numbers. A camera measures light, breaks it into tiny squares, and turns each square into a number. Those numbers are called pixels. The computer sees a giant grid of numbers — not a cat.
Let’s see like a computer: pixel coloring
Children color grids by number at three resolutions: 8×8 (64 pixels), 16×16 (256 pixels), 32×32 (1,024 pixels). The same cat image becomes progressively clearer. The computer still just sees numbers.
◆ Hands-on activity: Pixel Cat Children color a grid using a number key to reveal a cat — at 8×8, 16×16, and 32×32 resolution. The experience makes visceral the difference between seeing and calculating. Try the interactive version below.
Try it: color by number

8×8 — 64 pixels

Color each square using the key. What appears?

32×32 — 1,024 pixels

The same cat, more detail. Still just numbers to a computer.

Picture books as discussion openers
Eye Spy: Wild Ways Animals See the WorldGuillaume Duprat — each animal’s visual world shown as it actually appears to them. Perfect for introducing the idea that “seeing” is interpretation, not just measurement.
Hidden Systems: Water, Electricity, the Internet, and the Secrets Behind the Systems We Use Every DayDan Nott — graphic novel showing invisible infrastructure. Grounds the discussion of how computers process information in familiar everyday systems.
And They All Saw a CatBrendan Wenzel — the same cat seen through the eyes of different creatures, each perceiving something entirely different. A perfect picture book anchor for the “what’s in a cat?” discussion about perception vs. calculation.
3
Patterns & Predictions
Unit 3 · January 2026
▼

Unit 3 introduces the heart of how AI works: pattern recognition. Children explore patterns in art, in nature, in their own daily lives — then confront what happens when a computer learns patterns from incomplete or biased data. The Playground Scenario asks children to reason about fairness, missing information, and what it means to predict rather than understand.

What patterns do you notice daily?
Opening discussion — children name patterns from their own lives. Patterns in nature, in schedules, in people’s behavior.
Patterns, everywhere: OuiSi activity
Using OuiSi (Games of Visual Connection), children find unexpected connections across very different images — building intuition for how pattern recognition works before computers are mentioned.
Let’s make some rules
Using Which One Doesn’t Belong?, children sort shapes and articulate their own classification rules. Then a volunteer “robot” tries to guess the pattern — and often gets it wrong.
The Playground Pattern Scenario
A computer watches playground videos to predict which children will enjoy playing together. It learned that boys play soccer and girls jump rope. Discussion questions: What information does the computer actually receive? Whose experiences might be missing from the videos? When does noticing a pattern become assuming it always happens? What would it take for the computer to change its mind?
Let’s map how we made the decision
Children map their own reasoning as a decision tree — visualizing how AI systems structure choices and where assumptions enter.
Let’s teach a computer (try at home)
Using Google’s Teachable Machine: can you teach the computer to recognize a red object and a purple one? Bring your results to the next session.
◆ Ethical reasoning: The Playground Scenario Six discussion layers: Understanding the system → Pattern formation → Missing information → Generalization & assumptions → Comparing humans & computers → Revising predictions. Advanced extension: How confident do predictions sound even when based on limited examples? What clues could help us question a prediction?
Picture books as discussion openers
OuiSi: Games of Visual ConnectionUsed directly in the session — children find unexpected connections across images, building intuition for pattern recognition before the computer concept is introduced.
Which One Doesn’t Belong?: Playing with ShapesChristopher Danielson — used for the rule-making activity. Every shape in every group “doesn’t belong” for some reason. Children discover that classification rules depend on what you choose to notice.
Facts vs. Opinions vs. RobotsMichael Rex — introduces the difference between facts, opinions, and automated conclusions. A natural bridge to the Playground Scenario discussion about predictions vs. understanding.

Want to bring Kids Thinking Club to your school?

The full curriculum covers AI literacy, ethics, how machines learn, creative AI, and what it means to be a thoughtful human in an AI world. It’s designed to grow with children across grade levels — from first exposure to deeper reasoning.

Get in touch → ivi.kolasi@gmail.com
Berkeley, CA · Available for schools, libraries, and after-school programs