Ask first
Listen to what the learner already thinks. Avoid correcting immediately. Encourage simple explanations.
← AI AcademyFor families and educators
AI Discoverers · Ages 7–9
Explore together. Notice what children understand.
Use this simple guide to facilitate, discuss and review the eight introductory AI lessons. Its goal is to discover what learners understand, not to grade or compare children.
Family betaAdult-guidedPrintableNo child information collected
Explore session promptsTo print, use the Print command in your browser menu.
Before a session
Make room for curiosity.
Allow around 10–15 minutes per lesson. Invite children to predict, try, observe, and explain. Use the paper alternative whenever a screen is not suitable.
Try and change
Make one change at a time. Ask what happened and why. Treat mistakes as useful clues.
Explain together
Ask the learner to explain the idea in their own words or with a drawing. Repeat later if needed.
Eight session cards
One practical conversation for every lesson.
Choose the lesson relevant to the learner. Questions are examples, not a standardized test.
Lesson 01 / 08
Meet AI
Learning goal: Distinguish fixed rules, learned patterns and cases where the information is insufficient.
Ask before: Is every robot an AI robot?
Notice during: Ask for one example that follows fixed rules and one that learns from examples.
Ask after: Why is a beeping gadget not enough evidence to say it uses AI?
Misconception to watch for: A device can appear clever without learning.
Offline alternative: Sort household gadgets into fixed rules, AI possibilities or not sure.
Open lesson 01 →Lesson 02 / 08
Be the Robot
Learning goal: Describe, test and debug an ordered sequence of instructions.
Ask before: What happens if a robot misses one instruction?
Notice during: Have the learner build and follow directions one step at a time.
Ask after: What changed when the robot hit a rock?
Misconception to watch for: Following fixed instructions does not itself mean AI.
Offline alternative: Guide a partner across a paper four-by-four grid with arrow cards.
Open lesson 02 →Lesson 03 / 08
Teach the Robot
Learning goal: Observe how example choice changes a toy classifier's result.
Ask before: Would a green apple always be recognized if only red apples were shown?
Notice during: Try a green apple with color-only clues, then include shape.
Ask after: Why is it important to test on unseen examples?
Misconception to watch for: Getting a training example right is not proof of general learning.
Offline alternative: Sort drawn fruit by color and shape; introduce a surprising example.
Open lesson 03 →Lesson 04 / 08
Can Machines See?
Learning goal: Describe pixels and why fewer details can make images harder to interpret.
Ask before: What is a digital picture made from?
Notice during: Compare coarse and detailed pictures and invite two guesses.
Ask after: Why might a blurry picture be difficult to recognize?
Misconception to watch for: A pixel drawing is not a working computer vision system.
Offline alternative: Draw a secret object on squared paper, then remove details.
Open lesson 04 →Lesson 05 / 08
Guess the Next Word
Learning goal: Understand that different sentence contexts change plausible next words.
Ask before: Can more than one word fit a sentence?
Notice during: Compare classroom word weights for two different sentence starts.
Ask after: Does a plausible prediction guarantee the sentence is true?
Misconception to watch for: Illustrative word weights are not probabilities from a real language model.
Offline alternative: Play sentence-completion with three different beginnings.
Open lesson 05 →Lesson 06 / 08
AI Truth Detective
Learning goal: Distinguish confident statements from independently checkable evidence.
Ask before: Can someone sound certain and still be mistaken?
Notice during: Ask the learner to explain why a reference is more useful than a meme.
Ask after: How would you check a surprising science fact?
Misconception to watch for: Confidence or popularity is not evidence.
Offline alternative: Find one fact in a children's science book and say how it was checked.
Open lesson 06 →Lesson 07 / 08
Fair AI
Learning goal: Recognize uneven performance and the need for more testing.
Ask before: Should a helper work well for different people?
Notice during: Compare hypothetical correct-count bars for two speaking styles.
Ask after: Why can a small experiment not prove overall fairness?
Misconception to watch for: Better numbers in one test do not guarantee fairness everywhere.
Offline alternative: Create paper groups of examples and compare mistakes without recording anyone.
Open lesson 07 →Lesson 08 / 08
Invent a Helpful AI
Learning goal: Start with a real problem, compare simpler solutions and plan human oversight.
Ask before: Does every helpful invention need AI?
Notice during: Ask the learner to compare AI and non-AI approaches in the design card.
Ask after: Who checks a suggestion when it might be wrong?
Misconception to watch for: Choosing AI is not automatically better than a simple tool.
Offline alternative: Draw a useful helper and label its inputs, possible mistake and human check.
Open lesson 08 →After learning
Use this simple, paper-based observation checklist.
Tick observations on a printed copy. No online feedback form, learner profile, score or data upload is part of this guide.
For each item, mark Not yet, With support or Independently on paper, based on what the learner can explain.
- The learner describes the main idea in their own words.
- The learner predicts what might happen before trying an activity.
- The learner notices a mistake, uncertainty or unexpected result.
- The learner can change one example or instruction and explain the effect.
- The learner names a reason to question or check an AI output.
- The learner knows that a person stays responsible for important decisions.
Facilitator reflection: Which explanation was confusing? Which activity engaged the learner? What should be improved before using it again?
Privacy and safety
Protect learners while exploring AI.
Do not enter child names, addresses, photographs, voice recordings, school details or assessment notes into third-party AI tools. The Daneshkhana activities use teaching simulations, not real AI services. Discuss sensitive or confusing claims with a trusted adult.
This is a family beta support guide, not a validated examination, medical assessment or individual learning diagnosis.
What happens to feedback?
This page does not submit or store it. Keep observations on paper or use them privately to improve how you facilitate learning. Do not share personal records about children.
Back to AI Discoverers