Hacks on, hacks off: How one coding dojo is preparing kids for the AI world

As AI tools have advanced, anyone can now open ChatGPT to “vibe code” their way to a rudimentary app or generate an image for an online misinformation campaign—which often leaves kids the difficult task of separating what’s real from the AI slop.

“Will kids use AI to help facilitate writing their code? Absolutely. But at the end of the day, they’re still in control of it; it’s not controlling them.”

A coding “dojo” in Ottawa’s west end has launched an AI curriculum to train its young grasshoppers to do just that. Just a few blocks away from Canada’s largest technology park, Code Ninjas Kanata is one of 65 franchise locations teaching AI literacy to kids between five and 14 years old across Canada. Launched in September, the curriculum aims to teach young coders about AI safety and how it works under the hood, while also teaching them the skills to code without using AI.

“It’s important for us as a society to kind of teach the kids how [AI] works and how to use it responsibly,” Code Ninjas Kanata co-owner Ian Chan told BetaKit while donning a “Code Sensei” hoodie in November. “Here’s useful things that you can do with it, as opposed to just random things that kids will type into all these AI bots.”

Standing just a few feet behind his wife and fellow co-owner, Shuana Chan, Ian spoke to BetaKit in a hallway lit with paper lanterns. The hallway connects Code Ninjas’ two classrooms. One closer to the front desk sports nine seats around a central table, backgrounded by a whiteboard that reads “juniors” in jovial bubble letters, while the larger room in the back is meant for the older kids.

Kids typically attend Code Ninjas after school or on Saturdays to earn their belts by completing modules that teach them about coding through video game development. When earning their white belt, young kids learn to create a pixel-art sprite and how changing Y and X values can change the sprite’s position on the screen. As they grow older and learn more about coding in their quest for a black belt, the kids eventually learn new programming languages like Python, C#, Java, Lua, and Godot.

But why bother with all this when anyone can just ask Claude to make an app? Shauna said it’s important for kids to know the foundations of coding before they’re given AI to perform more complex functions.

“Will kids use AI to help facilitate writing their code? Absolutely,” Shuana said. “But at the end of the day, they’re still in control of it; it’s not controlling them.”

The Code Ninjas learning platform is a closed environment, meaning kids can’t just cheat their way to a black belt by copying and pasting code from another tab, Shauna explained. However, Code Ninjas has introduced a way for some kids to familiarize themselves with consulting an AI chatbot to find a bug in their code.

“It makes them a little more self-reliant, but it’s still super important that they don’t lose that learning moment,” Shauna said. “That’s where we come in, we facilitate the learning journey.”

Four kids were in Code Ninjas’ large classroom when BetaKit visited, working on their projects and giggling while playtesting the games that they and their peers had made.

When one of the children started playing Google’s Quick, Draw!, a tool in Code Ninjas’ AI curriculum, others followed. The game delivers a prompt, like “tree,” for the child to doodle while the AI attempts to guess what is being drawn. Afterwards, the game shows the database of doodles it was working with to identify that drawing as a tree.

All images courtesy Alex Riehl for BetaKit

Ian said this is meant to show the kids how AI models learn, and that it can take “100 iterations” for an AI to know what something looks like. He pointed out to a student that their attempt at drawing a tree resembled a drawing of a mouse in the database.

“This is what a tree looks like to him … it,” Ian corrected himself.

Code Ninjas also wants to teach its students to be critical in an AI world. Having been accused by the kids that an image of him wearing his black belt in taekwondo was generated by AI, Ian explained that kids these days already question everything; they just need to refine that instinct.

One module mixes in the iconic house hippo, a platypus, and many other real and artificial intelligence-generated images to help children practice their skepticism and explain to instructors why they think something is real or fake. Another provides a list of reliable sources for children to fact-check with.

As more kids and teens adopt AI, Hong Le, the mother of a Code Ninjas student, said she doesn’t have any concerns about the technology right now. Perhaps she will in her daughter’s future, she added, but she does believe in kids using their “critical thinking skills.”

“I feel like sometimes AI kind of gives them the answer, and they’re not using their thinking skills,” Le said, adding, “I think it’s something we all have to move towards, so it’s just adapting to it.”

Feature image courtesy Alex Riehl for BetaKit.