How to "Teach" AI to Teenagers

Led by the White House, AI has become the new “skill” for students and workers, and the U.S. is about to spend hundreds of millions “teaching AI” in schools.
But if we want a workforce ready for the future—here’s a radically different model that is already working.

This year alone, teenagers at Hack Club have built nearly 17,000 open-source technical projects—most using AI. That’s more teenagers building technical projects at Hack Club than taking the AP Computer Science in 30 U.S. states combined.

Our goal is 100,000 teenagers will be building technical projects with AI tooling by end of next year.

Building projects is the best way to develop AI fluency as well as a founders’ mindset. Teen projects range from simple video games and websites that require foundational CS skills, to increasingly complicated projects like custom operating systems, mobile apps, flight simulators and circuit boards.

At Hack Club, teenagers compete in building projects to earn prizes, and along the way learn “AI fluency.” For example, to code faster, they learn sophisticated AI tooling like GitHub Copilot, ChatGPT, Claude, Cursor, and Windsurf that we provide for free.

To do more with their projects, they might go deeper, experimenting with the machinery that underpins AI technology, working with GPUs, algorithms, and neural networks. AI ethics are important, and they wrestle in real time with navigating lines around what is original work vs copying– where AI is valuable and where human judgement and creativity is still irreplaceable.

Getting 100,000 teenagers to build projects in their free time is less hard than many think if you’re willing to get out of the preindustrial, 19th century mindset of how education is supposed to look. Hack Club runs big, stylized online competitions where tens of thousands of teens have been earning prizes for completing projects. (We’ve distributed $1m in prizes this year).

We also throw huge events in hundreds of cities, from software competitions to video game jams to Outward-bound-style challenges, and more than 10,000 teens have completed projects to qualify for hackathons, meeting friends along the way.

Any successful effort to prepare students and workers for the future must recognize that AI isn’t a fixed subject like algebra; it’s a dynamic set of evolving tools that require agency, experimentation, and freedom to master. And it must invest in motivation, because rigor and hard work is required to learn the foundational technical skills for mastering AI (familiarity with data, platforms, statistics, infrastructure, servers, files, and code).

We are moving from an economic model of 1 founder and thousands of workers, to thousands of founders with 1 AI tool. Time for the education model to change just as radically.