Colearnwith The Lossless Group

AI Explorers: AI Camp Syllabus for Ages 8–11

AI Explorers: Ages 8–11

The pitch for campers: Computers can learn, but only from what you show them. This week you'll teach one to recognise your stuff, build a robot sidekick that helps with something you care about, and find out what AI is great at and where it gets things hilariously wrong.

Part of AI Camp. Shared principles, the Agent Card, the Four Lenses rubric and the tools-by-age table live there.

Why this age band is built differently

  • No personal chatbot accounts. ChatGPT is not meant for children under 13. [1] Claude.ai requires users to be 18 or older. [2] Gemini can be enabled for under-13s only by a parent through Family Link. [3] So in this band, the instructor drives any chatbot on the big screen and the campers steer by voting, suggesting and fact-checking.

  • Training beats chatting at this age. Kids learn more about how AI works by training a model than by talking to one. AI4K12's guidance to educators is to use transparent demos, help students build mental models, and have them build AI applications. [4] Teachable Machine lets campers train image, sound and pose models in minutes, with no code. [5] Machine Learning for Kids plugs models they train into Scratch games. [6]

  • UNESCO level: Understand, with a taste of Apply. [7]

  • Play is the method. Every session follows Resnick's creative learning spiral: imagine, create, play, share, reflect, and imagine again. [8]

Ratio: at most 1 adult to 8 campers. Devices: Chromebook or laptop with webcam. Data rule: models are trained on objects, drawings, sounds and poses, never on faces. Nothing is saved after camp unless a family opts in.

Course 1: My Robot Sidekick

Maps to the original AI Mastery for Living Life.

Big question: If you had a robot sidekick, what would you want it to help you with, and what should it never do?

Campers will be able to:

  • Explain in their own words that AI learns patterns from examples (AI4K12 Big Idea 3: Learning). [4]

  • Train and test a classifier, and show how better examples make a better model.

  • Fill in an Agent Card (Goal, Tools, Memory, Rules, Check-in) for a sidekick in each of three life areas: me, my people, my school.

  • Name one thing their sidekick gets wrong and one way to catch it.

DayThemeWhat happens
1Is it a robot or a person?Unplugged game: campers act as a "robot" that can only follow exact instructions to make a sandwich. Then smart-speaker and chatbot demos on the big screen, where campers vote on whether each answer is right. They learn the word pattern and draw their dream sidekick.
2Teach it to seeIn Teachable Machine, campers train a "Sort My Stuff" model on three classes (say, healthy snack, treat, not food) using real objects. They break it on purpose with weird angles and bad lighting, then add better examples and watch it improve. This walks through the CSTA cycle: question, data, train, test, improve. [9]
3Sidekick for Me (personal)Each camper picks a personal goal: food, moving more, reading, or feeling calm. They fill in the Agent Card on paper first (brain first, then the bot). A movement sidekick is checked against the real guideline of 60 active minutes a day. [10] Then they build it in Scratch: a pose model that counts jumping jacks, or a sound model that cheers when you read aloud.
4Sidekick for My People and My School (social + school)Pairs build a "kindness reminder" bot for family or friends (a Scratch sprite that suggests one nice thing to do today) and a "homework helper" that quizzes rather than answers. The rule they test: does it make me do the thinking?
5Sidekick ShowcaseCampers polish one sidekick and present its Agent Card to families in three minutes: what it does, what it gets wrong, and when it asks a grown-up.

Capstone: one working Scratch or Teachable Machine sidekick plus three paper Agent Cards (me, my people, my school).

Course 2: Invent-o-Rama

Maps to the original AI Mastery for the Work/School World (Tech, Arts, Biz).

Big question: What could you invent if a computer could learn from you?

Campers rotate through three pods on Days 1–3, then pick one for their invention.

Campers will be able to:

  • Describe how computers sense the world (AI4K12 Big Idea 1: Perception) and how they learn (Big Idea 3). [4]

  • Make one thing in each pod: something that senses (Tech), something that creates (Arts), something that helps others (Biz).

  • Pitch an invention in under a minute.

DayPodWhat happens
1Tech pod: Robot ReflexesTrain a gesture model (thumbs up, wave, fist) and use it to steer a Scratch character or a classroom robot. The engineering design loop: plan, build, test, fix.
2Arts pod: Sound StudioTrain a sound model on claps, snaps and whistles that triggers instruments in Scratch, so you can play a song with your body. Discussion: if a computer helps make art, who's the artist? (Short answer: the person whose ideas and choices shaped it.)
3Biz pod: Helper StartupLook at the UN's 17 Sustainable Development Goals and pick one to help with locally, such as clean water, no hunger or life on land. [11] Design a helper bot, say a "Which Bin?" recycling sorter trained in Teachable Machine. Make a poster and a price ("it's free for our school!").
4Invention DayCampers choose a pod and build their own invention with a partner. Instructors push one upgrade: more training data, a second class, or a new way to use it.
5Invent-o-Rama FairScience-fair-style showcase. Families walk the room and try each invention. Each team explains one way their AI could be unfair, for example if it only learned from one kind of example.

Capstone: one working invention plus a one-page "Inventor's Notebook" with the problem, the examples used to train it, and how it was tested.

Course 3: Healthy Me, Happy Crew

Maps to the original AI Mastery for Personal & Social Life.

Big question: How can technology help you feel good and be a good friend, and when should you put it down?

Why it matters: Close relationships are among the strongest predictors of a long, healthy, happy life, according to the 80-year Harvard Study of Adult Development. [12] The US Surgeon General has called loneliness an epidemic. [13] So this course treats tech as a bridge to people, never a replacement.

Campers will be able to:

  • Explain why an AI can't really be your friend, even when it sounds friendly (AI4K12 Big Idea 4: Natural Interaction). [4]

  • Build a habit helper with a real health target.

  • Make a "put it down" rule for their own screen time.

DayThemeWhat happens
1Does the robot have feelings?Watch a chatbot on the big screen say "I'm happy to help!" Is it happy? Sort cards into "people can," "computers can," and "both, sort of." Introduce the idea that sounding human isn't the same as being human.
2Body botBuild a "Move Meter" in Scratch with a pose model: it counts movement minutes toward the CDC's 60-minute goal. [10] Talk about why the bot should cheer, not nag.
3Mind botBuild a "Calm Corner" helper: a breathing-pace animation, plus a feelings check-in made of buttons, not a chatbot. Its Rules box says: "If someone feels really bad, tell a grown-up."
4Crew botBuild a "Friend Quest" generator: Scratch picks a random kind, real-world mission ("Ask someone at lunch about their favourite game"). Campers do one mission at camp and report back.
5Family Showcase + Tech PromiseCampers demo a bot, then co-write a simple family tech promise inspired by the AAP Family Media Plan, ready to finish at home. [14]

Capstone: one helper bot plus a signed "Tech Promise" draft to take home.

Assessment for this band

Use the shared Four Lenses rubric from Program Design, scored mostly through conversation and showcase rather than writing. Look for:

  • Purpose: "I made this because…"

  • Build: it runs and a parent can use it.

  • Judgment: the camper can show one mistake the model makes.

  • Humanity: the Check-in box names a real person.

Instructor prep

  • Free lesson banks to pull from: MIT RAISE / Day of AI elementary lessons. [15]

  • Practise training and exporting a Teachable Machine model into Scratch through Machine Learning for Kids before Day 1. [6]

  • Prepare a physical "training kit" (snacks, recycling items, picture cards) so no camper has to photograph themselves.

Sources

Sources

  1. [1]
    Is ChatGPT safe for all ages?help.openai.com
  2. [2]
    Protecting the wellbeing of our usersanthropic.com
  3. [3]
    Guide your child's Gemini Apps experiencesupport.google.com
  4. [4]
    AI4K12: Sparking Curiosity in AIai4k12.org
  5. [5]
    Teachable Machineteachablemachine.withgoogle.com
  6. [6]
    Machine Learning for Kidsmachinelearningforkids.co.uk
  7. [7]
    AI competency framework for studentsunesco.org
  8. [8]
    Give P's a Chance: Projects, Peers, Passion, Playweb.media.mit.edu
  9. [9]
    AI Learning Priorities for All K-12 Studentscsteachers.org
  10. [10]
    Physical Activity Guidelines for School-Aged Children and Adolescentscdc.gov
  11. [11]
    The 17 Goalssdgs.un.org
  12. [12]
    Good genes are nice, but joy is betternews.harvard.edu
  13. [13]
    Our Epidemic of Loneliness and Isolation: The U.S. Surgeon General's Advisory on the Healing Effects of Social Connection and Communityhhs.gov
  14. [14]
    How to Make a Family Media Planhealthychildren.org
  15. [15]
    Resourcesraise.mit.edu

From client-content/Colearn/AI-Camp/Syllabi/AI Explorers - Ages 8-11.md · lossless-group/lossless-client-content