TL;DR: Ten years ago we talked about flipping education. Now the world is talking about AI literacy: UNESCO has released a student AI competency framework, and countries are packaging AI literacy into courses delivered by the credit hour. That is a good thing, and it has limits. An education experiment from a decade ago has a question about it: is there a syllabus for collaborative literacy? Can systems thinking be put on a timetable? These capacities only grow inside real tasks. The flip opened possibilities; what kids still need to climb in the AI era is figuring out how to actually grow competency, not just deliver it.

Ten years ago, the hottest word in education was “flip.” Turn teacher lectures into student self-study. Turn textbooks into projects. Turn classrooms into the world. That was when I started walking a non-conventional education path with kids.

Ten years later, the hottest phrase is “AI literacy.” And this time, the world moved faster.

Countries Are Scheduling AI Literacy into the Timetable

In 2024, UNESCO released an AI competency framework for students, breaking AI literacy into 12 competencies across four dimensions: human-centered mindset, AI ethics, AI technology and application, and AI system design. Its emphasis is on critical engagement, not stopping at the level of operating tools.

The framework was followed by action. Some countries added AI general education courses to primary and secondary school, requiring a minimum number of credit hours per school year, covering the full range from elementary through high school. Others took a national-team approach, setting targets to build a complete cross-level AI education system by 2030. Behind all of these moves is a shared logic: literacy can be curriculum-ized, standardized, and delivered by the hour.

That direction is reasonable. Curriculum is the only path that scales. It keeps AI education from being a privilege reserved for schools with ample resources, and it can reach remote and disadvantaged communities. If a country needs millions of children to have access to AI education, there is almost no other way.

But the ten-year education experiment I was part of left me with a different understanding, something I feel more than I can fully articulate.

Is There a Syllabus for Collaborative Literacy? Can Systems Thinking Be Timetabled?

Here is what my experience taught me: some capacities cannot be scheduled.

Judgment, collaboration, systems thinking, the things named in every future-skills report, do not come from completing a unit or passing an assessment. They need context, friction, and a shared goal that genuinely has to be met. In “Four Kids, One Summer, One Website That Had to Launch,” I wrote about how students really learn systems thinking: stuck in front of a blank page for two weeks, forced by an actual business problem to think it through. No single lesson could hand them that clarity. It had to be reached.

Collaboration works the same way. You can teach “five principles of teamwork” in a classroom and students will remember them. But a student learns to collaborate inside a real project, after arguing with someone who sees things differently, after restructuring the work division, after someone wants to quit and everyone pushes through anyway, and after finishing together. That capacity is not taught into existence. It is pressed into existence.

So when I see a design like “8 classes of AI literacy per school year,” I admire the ambition and feel the worry at the same time. A classroom can deliver knowledge. The rare practical capacities, those it cannot deliver.

This Is the Old Scale-vs.-Depth Problem in New Clothes

Task-based models have their own limits.

Real tasks depend heavily on family and adult resources. In “The Hidden Ledger of Education Innovation,” I worked through the numbers: one summer of real-task project work required more than 140 hours of adult involvement behind the scenes. That kind of intensity is not something every family or every school can sustain. If we only advocate for task-based learning and dismiss curriculum, we end up making good education a privilege for the few.

So curriculum-based and task-based approaches are two sides of the same old problem: scale or depth? Curriculum reaches scale. Tasks reach depth. The hardest thing in AI-era education is not choosing a side; it is figuring out how to connect curriculum’s scale with the depth of real tasks, so that more children who can now access AI education also have a genuine chance to grow their capacities inside something real.

The Flip Opened a Door. Climbing Is the Work on the Other Side.

Back to the two words I started with. Ten years ago I talked about flipping. Now I want to talk about climbing.

Flipping is opening a door. Children are no longer defined solely by a standard curriculum, and AI has made knowledge and answers cheaper than at any point in history. That is genuinely good. It is a door that has been opened.

Climbing is what happens after. When answers are freely available, what becomes scarce changes: the ability to take a vague problem and think it through, to work alongside people with different expertise, to bring coherence out of confusion and uncertainty, to take responsibility for your own choices. AI cannot replace those things yet, and they are exactly what real tasks grow.

So what do kids need to climb in the age of AI? More than learning to use the tools, tools will keep changing. It is growing the capacities that AI will not devalue: judgment, collaboration, accountability, and the drive to turn an idea into something real.

In the overview essay “From Flip to Climb,” I wrote that the heart of education is letting children meet the real world. AI hasn’t changed that: literacy can still only grow through practice. Techniques go out of date, but the literacy that truly takes root turns into character, stays with a child for a long time, and can shape a whole life.