AI literacy is not the ability to produce the longest prompt. It is the ability to decide whether AI belongs in a task, understand the limits of the output, protect people and information, and remain accountable for the result.
What AI literacy means
UNESCO’s AI competency framework for students organizes learning across four dimensions: a human-centred mindset, ethics, AI techniques and applications, and AI system design. It also describes a progression from understanding to applying and creating.
For a family, that framework can be translated into four everyday abilities:
- Understand: explain what a system is doing without treating it like magic or a person.
- Choose: decide when AI is appropriate and when a person, book, teacher, or experiment is better.
- Evaluate: check evidence, notice uncertainty, and look for missing perspectives.
- Act responsibly: protect privacy, respect authorship, follow school rules, and own the final decision.
The goal is calibrated trust: not believing every answer, and not rejecting every tool. A literate learner adjusts their level of trust to the evidence, stakes, and purpose.
Why this matters now
In a Pew Research Center survey conducted in fall 2025, 64% of U.S. teens ages 13–17 said they had used an AI chatbot. Fifty-seven percent had used one to search for information and 54% for schoolwork help. One in ten said chatbots help with all or most of their schoolwork.
These are survey findings about U.S. teens, not universal rates for every child or country. They also do not show that every use is frequent or harmful. They do show that many children are making AI decisions before families have agreed on shared language or rules.
UNICEF’s June 2026 multi-country snapshot found uneven AI adoption and uneven awareness of risk. It argues for both safer systems and more comprehensive AI-literacy support for children and caregivers.
Five conversations to try this week
1. “What do you think the system knows?”
Ask a child whether the answer came from a verified database, a prediction, a search result, or something else. The point is not to demand a technical lecture; it is to separate fluent language from evidence.
2. “What should never go in a chat?”
Practise identifying passwords, precise location, school schedules, private images, health details, financial information, and another person’s story. Use fictional names and minimum necessary detail in low-stakes examples.
3. “How could we check this?”
Ask what kind of source would provide the best evidence. A current government page may be appropriate for a public rule; a primary document for a historical claim; a teacher-approved text for an assignment; and a qualified professional for high-stakes advice.
4. “What part of this work is yours?”
Have the learner explain what they decided, wrote, verified, or changed. Authorship is not only about who typed the words; it is about who understands and can defend the result.
5. “When should we stop?”
Normalize closing the tool when an output becomes private, sexual, hateful, unsafe, manipulative, or emotionally serious. A child should be able to ask an adult for help without fearing punishment for revealing the conversation.
Use the Pause → Probe → Prove routine
What is the task? Is AI appropriate? What information is unnecessary? What could go wrong if the answer is inaccurate?
Ask: “What are you assuming?”, “What might be missing?”, “How certain are you?”, or “What evidence would change this answer?”
Verify important claims with a suitable source or qualified person. Keep the source and date when the information may change.
This routine can be used without an AI tool. Try it with a social post, advertisement, recommendation, or search result. The durable skill is evidence-aware judgment.
Adjust the conversation by age
Ages 8–11: concrete examples and shared use
Use familiar systems such as recommendations, autocomplete, or a game character. Operate tools together, avoid child accounts where possible, and focus on one rule at a time: “AI can make a useful guess, and we still check.”
Ages 12–14: sources, privacy, and school rules
Introduce hallucinations, training data, bias, and disclosure. Compare two outputs and ask which is better supported. Review each provider’s age requirements before allowing independent use.
Ages 15–17: independence with accountability
Discuss authorship, copyright, emotional use, persuasive design, and high-stakes decisions. Ask teens to propose the family or school rule themselves, then explain the trade-offs.
Age is only one factor. A child’s developmental needs, the tool’s design, the topic, and the stakes all matter. Emotional support, private images, or health advice require more adult involvement than generating practice questions.
Talk about schoolwork without shame
A child may turn to AI because a task is confusing, tedious, inaccessible, or high-stakes. Start by asking them to show the process. If the first response is accusation, the next AI use is more likely to become hidden.
A practical schoolwork protocol can allow different levels of assistance:
- Explain: ask for another explanation, then solve a new example independently.
- Brainstorm: generate possibilities, but choose and develop the idea personally.
- Feedback: request comments on clarity without accepting an automatic rewrite.
- Verify: check claims and citations in approved sources.
- Disclose: follow the teacher’s rules for acknowledging AI help.
If AI is prohibited, help the learner find another support: a teacher, librarian, peer, accessibility service, worked example, or tutoring resource.
A one-page family AI policy
Keep the policy short enough to remember. A useful version can fit on the refrigerator:
Our family uses AI to learn, explore, and create—not to hide work or replace people who are responsible for our wellbeing. We protect private information, verify important claims, follow school rules, and tell a trusted adult when a conversation feels unsafe or too personal.
Review the policy every few months. Tools, school expectations, and a child’s independence will change. The conversation should evolve with them.