AI Learning Roadmap by Age: 7–9, 10–12, and 13–14
A strong AI learning path starts with curiosity and safety, then moves from recognizing patterns to building, testing, and explaining small projects. Ages are guides, not limits. Choose the path that matches the learner's reading level, attention span, coding experience, and ability to question an AI answer.
Before starting: the family AI agreement
Agree that an adult will know which tools are used; private information will stay out of prompts; important claims will be checked; AI will not impersonate people or complete dishonest schoolwork; and the learner can stop and ask for help whenever an output feels confusing, upsetting, or unsafe.
Ages 7–9: explore patterns with an adult
Goal: understand that AI finds patterns in examples and can make mistakes. Keep sessions to about 15–25 minutes with an adult beside the learner.
Weeks 1–2: Sort picture cards into groups and discuss which clues a computer might use. Compare human rules with a simple computer rule.
Weeks 3–4: Use a supervised, no-personal-data image or sound classifier. Train it on safe household objects or newly made drawings.
Weeks 5–6: Ask an adult-operated chatbot three fact questions, then check the answers in a trusted book or primary source. Keep a correct, uncertain, and wrong chart.
Weeks 7–8: Create a short story with AI suggestions, but let the child choose characters, revise the ending, and explain which ideas are their own.
Milestone project: a Pattern Detective poster showing examples, rules, one mistake the model made, and what the learner changed.
Ages 10–12: build, test, and explain
Goal: move from using an AI tool to testing how inputs, examples, and instructions change results. Sessions can be about 30–45 minutes.
Weeks 1–2: Learn input, model, output, training example, prediction, and hallucination using child-friendly examples.
Weeks 3–4: Build a small Scratch project that reacts to a prediction or uses a rule-based character to compare programmed rules with learned patterns.
Weeks 5–6: Run the same prompt five times, record differences, and rewrite the prompt to make success criteria clearer.
Weeks 7–8: Compare an LLM with a vision-language model using a newly drawn diagram. Verify every visual observation.
Weeks 9–10: Complete a bias activity by checking whether the examples represent different cases fairly. Add missing examples and retest.
Milestone project: a Fact-Checking Challenge with the original claim, sources used, final judgment, and a reflection on why confidence is not evidence.
Ages 13–14: design responsible AI projects
Goal: plan a useful project, document evidence, evaluate failures, and communicate safety tradeoffs. Learners can use visual coding or beginner Python with adult awareness.
Weeks 1–2: Review machine learning basics, data quality, evaluation, privacy, bias, and the difference between a demonstration and a reliable product.
Weeks 3–4: Choose a narrow problem and write success criteria before selecting a tool. Avoid projects that identify people, infer emotions, or process private information.
Weeks 5–6: Build a small prototype, keep a test log, and record inputs, outputs, errors, time, cost, and unexpected behavior.
Weeks 7–8: Create edge cases designed to make the project fail. Improve instructions, examples, or interface boundaries and retest.
Weeks 9–10: Write a model card for the project: purpose, intended users, data used, known limits, privacy notes, and situations where it should not be used.
Milestone project: a documented AI demonstration with screenshots, test notes, at least one failure, one improvement, and a short responsible-use explanation.
How parents can measure progress
Ask the learner to explain what the system received, what it produced, how they checked it, what could go wrong, and what they would change. Progress is not the number of tools used. Progress is better questions, clearer evidence, safer choices, stronger explanations, and more independent thinking.
Recommended next steps
Start with one 8–10 week path, save the work in a simple portfolio, and repeat a favorite project with better testing. Families who want a guided introduction can request the free online AI class for ages 7–14.
Published and last materially reviewed: August 1, 2026. By AI Education for Kids Editorial Team.