Generative AI is no longer a future problem for education. It is already part of childhood.
The OECD reported in 2026 that three-quarters of students aged 16 and over used generative AI tools in 2025.
At the same time, the World Economic Forum says 39 percent of workers’ core skills are expected to change by 2030, with creative thinking, resilience, and curiosity rising in importance.
That combination should force a change in what schools reward, because the definition of useful intelligence is shifting in real time.
For generations, education has largely trained students to solve problems placed in front of them: calculate the answer, write the essay, complete the assignment, pass the test.
But when a machine can produce a plausible answer in seconds, reaching the answer can no longer be the highest form of intellectual achievement.
One of the most valuable skills may increasingly be identifying the problems worth solving.
By creating problems, I do not mean manufacturing difficulties. I mean noticing what others have overlooked, challenging the premise of a question, identifying a better question, and imposing useful constraints.
Innovation rarely begins with someone answering the obvious question faster. It begins when someone asks why everyone is answering that question in the first place.
This is also why the usual debate over whether AI makes young people more or less creative misses the point. AI can do both. It can flatten thinking when students accept its first response as authoritative.
A Microsoft Research study of 319 knowledge workers found that greater confidence in generative AI was associated with less critical-thinking effort.
But AI can also accelerate exploration, expose users to alternatives, and make experimentation dramatically cheaper.
The danger, then, is not that young people will have too many ideas. It is that they may never develop the judgment required to know which ideas are good.
That judgment is built through friction. A student who struggles with a question, tests an assumption, discovers that an approach fails, and tries again is not wasting time.
That student is developing taste, skepticism, and discernment. Those qualities are what allow an experienced person to look at an impressive AI-generated response and ask: Is this actually right? Is it useful? Is it original? What is missing?
The opportunity is equally significant. If AI can handle more routine retrieval, drafting, and iteration, education can spend more time on abilities historically squeezed to the margins: curiosity, experimentation, discussion, invention and decision-making.
The OECD’s 2026 Digital Education Outlook warns that general-purpose AI can improve task performance without necessarily improving learning, while AI used with deliberate pedagogical purpose can support critical thinking, creativity and collaboration.
This matters beyond classrooms. The same WEF jobs research shows that employers expect human skills such as creative thinking, resilience, and lifelong learning to grow in importance alongside AI and technological literacy.
The future will not belong simply to people who know how to operate AI. Nearly everyone will. Advantage will come from knowing what to ask of it, when to distrust it, and what to do after it gives you 10 plausible answers.
Schools should therefore change assessment, not merely add AI policies. Beginning with the 2027-28 academic year, education systems should require at least one “question-first” assessment in every secondary-school course each term.
Students should be graded on how they define the problem, challenge assumptions, generate competing possibilities, evaluate evidence, explain why they chose one path, and disclose how AI was used. The final answer should be only part of the grade.
If we continue educating children primarily to produce answers, we risk preparing them to compete with machines on precisely the kinds of tasks AI is improving at fastest.
If we teach them to frame better problems and make better judgments, AI can become something far more useful: not a substitute for creativity, but a tool that gives human creativity more room to matter,
In the age of artificial answers, the most human advantage may be knowing which questions are worth asking.
Or in other words, AI has the potential to make us more human, not less.
Dan Klitsner is an industrial designer, toy inventor, and creative innovator best known as the inventor of the Bop It and founder of Bop It for Good.