Artificial intelligence has already entered the classroom, though not always explicitly. Many students use these tools to complete assignments more quickly, to generate ideas, to manage their schedules, or simply to find answers. A recent study from the Brookings Institution indicates that more than ninety percent of teenagers in the United States report using such tools on a daily basis. This statistic, while specific to one country, signals a global shift in how younger generations engage with learning. My own experience in the classroom confirms this pattern, though with an instructive nuance. When I ask my students whether they employ artificial intelligence in their academic work, their first response is almost always denial. Yet, after a moment of hesitation, they admit not merely to occasional use, but to what they themselves describe as abuse—a reliance so pervasive that it has replaced rather than supplemented their own cognitive effort. The term is theirs, and it is telling. What changes, therefore, is not simply the presence of technology, but the student’s relationship with learning itself. Learning is not merely about arriving at an answer. It is about understanding that answer, questioning it, explaining it to others, and, on occasion, doubting it. When answers become instantaneous, and when the very frequency of their acquisition tips into dependency, the cognitive labour that once accompanied discovery risks disappearing. This is where the educational work continues. As educators and caregivers, we must decide what we will not delegate, not because the technology fails to function, but because certain skills develop only when students think for themselves.

If a student can obtain a well‑formulated answer in seconds, what role does thinking play in that process? The temptation to bypass struggle is understandable, but struggle is precisely where resilience is forged. Learning involves pausing, becoming confused, questioning assumptions, synthesising disparate pieces of information, reading closely, debating with peers, failing, and trying again. When every obstacle is resolved too quickly, the learner loses the very experience that allows for genuine understanding. That is why it remains crucial for students to learn to question information, to compare sources, and to construct their own ideas—even when a seemingly complete answer already sits before them. Having an answer is not the same as understanding it. A student who can recite a solution produced by an algorithm has not necessarily grasped the underlying reasoning, the assumptions embedded in the model, or the limitations of the data on which it was trained.

Creation takes time. It requires testing ideas, changing direction, making decisions, and, above all, making mistakes. If this process is shortened too much, what disappears is not merely the result, but the experience of thinking originally. When all answers are immediately available, exploration can be diminished. That is why it is essential to continue creating spaces where students must ask questions, not merely look for them. Spaces where failure is acceptable. Spaces where they can grow stronger through effort rather than through the avoidance of difficulty. As educators, we must resist the pressure to optimise every moment of learning for speed and efficiency. A curriculum that prizes only correct answers and rapid completion inadvertently teaches students that the journey is irrelevant. Yet the journey—the confusion, the false starts, the moment of insight—is precisely what builds intellectual resilience.

Much of learning happens with others: in conversation, in disagreement, in an idea built collaboratively through discussion and debate. This type of experience cannot be fully replaced by digital interaction, however sophisticated the technology. When students articulate their reasoning to a peer, defend a position against critique, or revise a shared project based on feedback, they engage in forms of thinking that no algorithm can replicate. These social dimensions of learning are not supplementary; they are central to the development of judgement, empathy, and intellectual humility. Today, more than ever, knowing how to evaluate information is essential—even more so than when the internet first flooded our classrooms. Not everything that sounds clear or authoritative is necessarily true. Students must learn to identify bias, to recognise the limits of automated responses, and to distinguish between plausible assertions and verifiable evidence. This skill, often called information literacy, becomes increasingly critical when generative artificial intelligence produces text that is fluent but not always accurate.

Using technology also involves making decisions about what is shared, and decision‑making implies responsibility. Students need to understand how data works and why it is important to protect their own information and that of others. The convenience of artificial intelligence should not come at the cost of carelessness. Educators have a role in fostering an ethical framework: one that asks not only “Can I use this tool?” but also “Should I? Under what conditions? With what transparency?” Technology can facilitate many things. It can accelerate research, organise ideas, and provide instant feedback. But thinking, understanding, and taking a stand are still processes that require time, guidance, and deliberate practice. And that—that cannot be delegated. To raise resilient learners in an artificial intelligence world, we must protect the spaces where struggle is permitted, where questions are valued over answers, where collaboration and critique happen face to face, and where students learn not only how to use tools, but when and why to trust their own judgement. That is the educational work that remains. And it has never been more urgent.