What Should Young People Aged 12–18 Learn About Artificial Intelligence?
Artificial intelligence is already changing the way young people learn. It can explain a subject, answer questions, suggest ideas, edit texts and produce a polished assignment within seconds. All of this offers opportunities to make learning easier. But it also raises an important question: When a task is complete, has the young person actually learned something, or have they simply produced something that looks good?
A ready-made answer does not mean that learning has taken place. Artificial intelligence can produce a polished assignment, a convincing explanation or a result that appears to be correct. But if the young person has not participated in the thinking process, checked the information or explained what they have learned in their own words, the finished product does not demonstrate genuine learning.
For this reason, we cannot approach AI literacy for young people aged 12–18 as simply knowing how to write effective prompts or use new tools. Young people need to understand how AI works, that it can make mistakes and that even highly convincing answers may not always be correct. They must also learn to decide when to use AI, which parts of a task they can delegate to it and how to evaluate its output.
Recognising the learning need before seeking support from AI
If a young person cannot identify what they already know, where they are struggling and what kind of support they need, they cannot use AI intentionally.
They may not understand a subject at all, be confused about one particular concept or understand the subject but struggle to organise their thoughts. Each of these situations requires a different kind of support. When young people cannot distinguish between them, it becomes very easy to ask AI to complete the entire task for them.
This is where AI literacy and metacognitive learning come together. AI literacy helps young people understand the tool in front of them, including its possibilities and limitations. Metacognitive learning enables them to manage their own thinking and learning while using that tool.
What is metacognitive learning?
In its simplest form, metacognitive learning means thinking about one’s own learning. Young people consider not only what they have learned, but also how they learned it, where they struggled and whether the method they used was effective.
When using AI, this skill can become visible through questions such as:
What am I expected to learn from this task?
Which parts of the subject do I understand, and where am I struggling?
Which parts should I do myself?
What kind of support can I ask AI to provide?
Did the support help me think, or did it think for me?
Have I verified the information using other sources?
Can I now explain what I have learned in my own words?
In this way, using AI is no longer a skill separate from the learning process. The young person does not simply accept an answer; they also consider the role that answer plays in their own learning.
Plan, monitor, evaluate
We can approach this way of thinking in three stages: planning, monitoring and evaluation.
During the planning stage, the young person first identifies their goal. They consider what they are trying to learn, what they already know about the subject and whether they genuinely need AI. They then decide which part of the task AI could help them with.
Instead of asking AI to write the entire assignment, for example, they could ask it to explain a difficult concept using a different example, show the difference between two approaches or provide feedback on a draft they have already prepared.
During the monitoring stage, they consider whether the support is actually helping. Is the information provided by AI correct? Is it based on a reliable source? Is it genuinely relevant to the question? Could the answer be missing an important perspective? Is the explanation helping them understand the subject, or merely allowing them to complete the task more quickly?
During the evaluation stage, they reflect on their learning at the end of the process. Can they explain the subject without repeating the AI’s wording? Can they explain why they accepted or rejected its suggestion? Would they know how to approach a similar question in the future?
At this stage, the real measure of success is not how polished the final text appears, but how well the young person has managed their own thinking process.
Asking these questions is not enough
Simply giving young people a list of questions will not automatically help them develop these skills. Telling them to “check the information,” “explain it in your own words” or “do not immediately trust AI” is important, but we also need to show them how to do these things.
At first, parents and educators can make the thinking and evaluation process visible by working through it with the young person:
“What claims are being made in this answer? Which ones do we need to verify? Has AI provided a source? Does that source actually exist? What perspectives might be missing from the response? Can we find the same information elsewhere? Now let’s reconsider the AI-generated answer in light of what we already know.”
The learning goal can also be identified together:
“Is the purpose of this task simply to submit a piece of writing, or is it to learn about a particular subject? Which part are you struggling with? Instead of asking AI for the answer, how could we ask it to help you?”
These conversations do more than tell young people what to do; they show them how decisions are made. Over time, young people begin asking themselves the same questions and forming their own judgments about AI-generated output.
Support should gradually decrease
Not every young person needs the same level of guidance. Their age, prior knowledge, motivation, digital experience and ability to manage their own learning will all affect how much support they need.
At first, an adult and a young person might set a learning goal together, examine an AI-generated answer and explore different ways to verify information. As the young person develops these skills, both adult guidance and the support provided by AI should gradually decrease.
Effective support should be like scaffolding that is removed over time, not a permanent structure. Success should not be measured by how often a young person uses AI, but by whether they can eventually manage independently the thinking and learning processes they initially carried out with support.
Too much support, even when well intentioned, can make it harder for young people to create their own plans, recognise where they are struggling and find solutions. Too little support can leave a young person who has not yet developed the necessary skills alone at the point where they need help. The goal, therefore, is neither to ban AI completely nor to encourage its use at every stage, but to adjust the level of support according to the young person’s needs.
AI should work with young people, not in their place
Our goal should be to raise young people who see AI not as a tool that thinks and creates for them, but as a form of support that accompanies them throughout the learning and creative process.
Young people can ask AI for ideas, examples, explanations or feedback. But they should not use that support as a substitute for their own thinking. They should use it to develop their ideas, identify gaps and evaluate what they have produced.
As parents and educators, our responsibility is to show young people how AI can support them and at what point they need to return to their own thinking.
The true measure of AI literacy is not simply whether a young person knows how to use the technology. It is also whether they know when to use it, what to delegate to it, how much to trust its answers and when to continue without it.
AI can support learning and creative work. But it should not learn, think or create in place of the young person.
Recommended programmes and websites for ages 12–18
🎓 Day of AI — MIT RAISE | Ages 11–18Explores how AI works alongside ethics, bias and safety.
🧠 Experience AI | Ages 11–14Uses practical activities to teach machine learning, data and bias.
💻 Code.org — How AI Works | Ages 12–18Explains the basic principles behind large language models, training data and algorithms.
⚖️ AI4ALL — AI & Ethics | High school levelEncourages young people to discuss the ethical effects of AI and consider the consequences of different decisions.
Sources
UNESCO — AI Competency Framework for Students (2024)
Presents four key dimensions for helping students use AI safely, meaningfully and responsibly: a human-centred mindset, ethics of AI, AI techniques and applications, and AI system design.
UNESCO — Guidance for Generative AI in Education and Research (2023)
Explores how generative AI can be used in education in ways that preserve human agency and learning goals while remaining age-appropriate and pedagogically meaningful.
Education Endowment Foundation — Metacognition and Self-Regulated Learning (2025)
Explains how students can develop the ability to plan, monitor and evaluate their learning, as well as how adults can model thinking processes and gradually reduce support.
OECD — Digital Education Outlook 2026 (2026)
Examines how generative AI can support learning when used in pursuit of clear learning goals, while warning that when it replaces cognitive effort, producing better-looking work does not necessarily lead to lasting learning.
Comments