THE MERIT BADGE · SCOUTING AMERICA
⚡ taught to you by an actual AI ⚡
Every word, every color, every game in this deck was generated by an AI — the same kind of technology this badge is about.
A human typed a few sentences describing what they wanted. The AI did the rest in a few minutes. By the end of today, you'll understand exactly how that's possible — and how to do it yourself.
16 key terms, decoded
Spot AI everywhere + play "AI or Not?"
Machines on autopilot
Play "What Would You Do?"
Spot the fakes, protect yourself
Prompt like a pro
Your AI project or lesson
Get paid to do this
COUNSELOR NOTE — This deck maps 1:1 to the official requirements (updated Oct 2025). Discussion requirements still happen out loud — the slides are your launchpad, not a substitute.
The games in this deck are played out loud, as a group. Wrong answers are welcome — that's literally how machine learning works too.
An AI made this deck. AIs make mistakes. If something seems off — say so. Catching AI errors is a skill this badge teaches.
You'll need your examples, timelines, and ethical guidelines for your counselor sign-off. Future you says thanks.
Every world has its own language. Gamers say "GG." Pilots say "roger." AI people say "machine learning." Level 1: crack the code.
The next four slides are flash cards — click any card to flip it and reveal the definition.
Pro move: before you flip, guess the definition out loud. Your counselor needs to hear you explain these in your own words anyway — this is your practice run.
Technology that lets computers do things that normally need human smarts — recognizing faces, understanding language, making decisions.
Teaching a computer by showing it tons of examples instead of writing exact rules. Show it 1M cat photos → it learns "cat." Like practice, but for software.
Machines doing repetitive tasks by following pre-set rules — no learning, no thinking. A dishwasher automates. It does not ponder.
Writing step-by-step instructions (code) that a computer follows exactly. Computers are obedient but extremely literal — they do what you say, not what you mean.
Programs that run tasks automatically — chatbots that answer questions, game bots, bots that scan the web. Some simple, some smart.
AI that doesn't just answer — it takes action toward a goal: planning steps, using tools, adjusting when things change. Like an NPC that actually improvises.
Software "coworkers" that handle whole office jobs end-to-end — processing invoices, answering support emails — like a virtual team member.
AI that's great at ONE thing — chess, playlists, face unlock — and useless at everything else. Every AI that exists today is narrow AI.
A future AI as flexible as a human — able to learn anything. Doesn't exist yet. Scientists debate if or when it will.
A theoretical AI smarter than all humans combined. Pure speculation today — but it's why serious people work on AI safety.
Information in any form — numbers, words, photos, clicks, songs, scores. Data is to AI what food is to you: it can't grow without it.
Organized digital filing cabinets where data lives so computers can find it fast. Your school stores grades in one. Games store your saves in one.
Labeled boxes in code that hold values that can change — like score = 250 in a game. Get a power-up, the variable updates.
TRY IT — Name three variables your favorite video game must be tracking right now. (Health? Ammo? XP?)
The event that kicks things off. WHEN the clock hits 6:30 → WHEN a customer emails → WHEN motion is detected...
The individual jobs that get done: send the reply, sound the alarm, turn on the light. One task = one action.
Trigger + tasks linked into a chain. WHEN order placed → charge card → print label → email tracking. That's a workflow.
Plot twist: you've been using AI since before breakfast. Level 2: learn to see the invisible machines all around you.
REQ 2a CHECK — That's 10 examples. Now make the list yours: which of these did YOU actually use today? Add any we missed.
Requirement 2d: done. Discuss with your counselor: which round fooled the most people, and why? The line between "smart" and "just automatic" is exactly what this badge is about.
Alan Turing asks: "Can machines think?" and proposes a game to find out.
Scientists at the Dartmouth workshop coin the term "artificial intelligence."
IBM's chess computer defeats world champion Garry Kasparov.
Siri ships on the iPhone; IBM Watson wins Jeopardy! against champions.
DeepMind's AI beats Go legend Lee Sedol with a move no human would play.
Chat AI reaches 100M users in 2 months — fastest-growing app in history. The AI era goes mainstream.
REQ 2e CHECK — Pick any FIVE milestones (these six or your own research) and sketch your personal timeline to show your counselor.
Before machines could think, they could repeat. Level 3: the 200-year-old technology hiding inside everything.
Follows the same pre-written rules every single time. Never learns, never improves, never surprises you.
IF temp < 68° → turn on heat.
Your thermostat. Every time. Forever.
Learns from experience, spots patterns, handles situations it's never seen before, and gets better over time.
Netflix noticing you binge heist shows in October → recommending one it thinks OCTOBER-you will like.
QUICK POLL — Automatic doors at Walmart: recipe or chef? (Recipe — a motion sensor follows one rule. No learning happening.)
Alarm clocks · thermostats · dishwashers · washing machines · toasters · automatic doors · sprinkler systems · streetlights on timers · vending machines · robot vacuums (bonus: modern ones add AI to map your house!)
Assembly-line robots welding cars · ATMs dispensing cash · self-checkout lanes · warehouse conveyor sorting · scheduled payroll runs paying everyone on the 1st
Bell schedules · scantron/auto-graded quizzes · attendance systems texting parents · library self-checkout · lunch account auto-reload
REQ 3a–c CHECK — That's 10 + 5 + 5. Your turn: swap in examples from YOUR house, YOUR school. Weirdest example wins.
Ask a human to fold 10,000 shirts and by #500 they're sloppy, by #2,000 they're miserable, by #5,000 they're making mistakes.
A machine folds shirt #10,000 exactly like shirt #1. That's the automation superpower: perfect consistency, zero fatigue, no human needed once it's set up.
Works 24/7. No sleep, no snack breaks, no scrolling.
Same steps, same result — fewer errors on repetitive work.
People move to the creative, judgment-call work machines can't do.
Edmund Cartwright's loom weaves cloth automatically — the Industrial Revolution accelerates.
Moving assembly line cuts the time to build a car from ~12 hours to ~90 minutes.
First industrial robot starts work at a GM plant, handling hot metal parts humans hated.
Cash machines automate the bank teller's most repetitive job — 24/7.
Home robots go mainstream — millions of floors vacuumed with zero human effort.
Warehouse robots, self-driving features, smart factories — the recipe and the chef team up.
REQ 3e CHECK — Choose five milestones for your own automation timeline. These work, or dig up your own (the dishwasher has a great origin story — look up Josephine Cochrane).
Halfway checkpoint: you can now speak AI, spot it in the wild, and tell a recipe from a chef.
With great compute comes great responsibility. Level 4: the questions that don't have an answer key.
AI learns from human data — including human unfairness. Facial recognition has misidentified people with darker skin far more often. Hiring AIs trained on old data have downgraded women's résumés.
garbage in → garbage out
AI is hungry for data — YOUR data. Your location, your face, your voice, your chats. Who collected it? Who's it shared with? Once it's out, you can't take it back.
if it's free, you're the data
Should an algorithm decide who gets a loan? A job interview? Parole? When AI gets it wrong, who's responsible — and can you even appeal to a human?
who watches the algorithms?
Five scenarios, five debates. Notice a pattern? The Scout Law kept showing up: trustworthy, helpful, kind, brave. Ethics in AI isn't new rules — it's old values applied to new tech.
Companies have AI policies. Countries have AI laws. Now it's your turn: write 3–5 rules YOU will follow when using AI.
Steal from the Scout Law. It was written in 1908 and somehow predicted every AI ethics debate of the 2020s.
Alan Turing's idea: put a judge in a text chat with a hidden human and a hidden computer. The judge can ask anything. If the judge can't reliably tell which is which, the machine passes.
Turing's real point: instead of arguing about whether machines "think," test whether they can convince us they do.
Seeing used to be believing. Level 5: the badge's most serious level — because this one can happen to you or your friends.
Deepfakes use AI to create video, photos, or audio of someone doing or saying things they never did. A cloned voice calling your grandma. A fake video of a celebrity "endorsing" a scam. A fabricated image of a classmate.
Don't forward it, even to "warn people." Every share trains the algorithm to spread it further.
Screenshot the post, the account, the URL, the date — before it disappears. You'll need proof.
Parent, teacher, Scout leader. This is not a "handle it yourself" situation — that's brave, not weak.
Report on the platform. If a fake image targets someone under 18, NCMEC's free Take It Down service (takeitdown.ncmec.org) helps get it removed. Serious cases → school & police.
DISCUSS — Why is "don't share" step one? What does the Scout Oath's "help other people at all times" look like in a group chat where a fake is spreading?
Talking to AI is a skill — and most adults are bad at it. Level 6: become the person your family asks for AI help.
Modern AI reads mountains of examples — text, images, sound — and learns the patterns. Then it predicts: "given everything I've seen, what most likely comes next?"
That's why this deck exists: the AI that wrote it learned patterns from millions of documents, slides, and websites — then predicted what a great merit badge deck would look like.
AI can state totally made-up "facts" with total confidence. Always verify anything important.
It only knows what it was trained on — recent events may be missing unless it can search.
Trained on human data = trained on human blind spots.
It predicts patterns brilliantly — but it has no lived experience, feelings, or common sense guarantees.
"Write about dogs" ❌
"Write 100 funny words about why pugs would fail a PE class" ✅
"Act as a soccer coach explaining offsides to a 10-year-old." Roles focus the answer.
"Here's a flashcard I like — make 10 more in this exact style." AI copies patterns beautifully.
"Give me a table," "keep it under 50 words," "make it a checklist." You get what you specify.
First answer's a draft, not a verdict. "Make it funnier." "Simpler." "Now for a 5th grader." Follow up!
Using all five on purpose. Congrats — that's a real job skill now.
REQ 6d CHECK — Now write YOUR three: clear AI instructions for three school tasks (study guide, essay outline, explaining a math problem...). Use the 5 moves. Demo them for your counselor.
One Scout becomes the robot. Others give instructions to make a PB&J (or tie a knot, or fold a flag). The robot does exactly what's said — nothing more.
"Put peanut butter on the bread" → robot places the sealed jar on top of the bag of bread. As instructed. 😈
Enough theory. Level 7: the requirement where you actually make a thing. Pick your path.
Solve something YOU care about, then walk your counselor through your goals, your data, and your ethics plan.
Design and run a lesson for your patrol — explanations, examples, and a hands-on demo.
Final level: the jobs. Most of them didn't exist when your parents were your age. Some don't exist yet — you might invent them.
Builds and trains the AI models themselves.
~$130–170k · CS/math degree common
Finds the stories hiding in mountains of data.
~$110–140k · stats + coding
Gives AI a body — drones, rovers, factory arms.
~$100–130k · engineering degree
Keeps AI fair, safe, and legal. Level 4, as a career.
~$90–150k · law, philosophy, or CS
Teaches, tests, and steers AI systems for companies.
varies widely · newest job here
Pick 3 careers → deep-dive ONE: training, education, certifications, salary, job outlook. (Salaries here are rough US figures — verify current numbers in your research!)
Parents' coworkers, local college professors, engineers at nearby companies — people love talking about their jobs to curious Scouts. Video calls count.
Fun fact: you technically interviewed an AI today. It made this deck. But requirement 8b wants a human — they have better stories.
16 terms — explain them to your counselor in your own words
Examples ✓ AI-or-Not ✓ Timeline: yours to sketch
Examples ✓ Repetitive tasks ✓ Timeline: yours to sketch
Big 3 concerns ✓ Scenarios ✓ Write your AI code + Turing test
What they are + the 4-step response playbook
5 moves ✓ Write your 3 school-task prompts
Homework: your project or your lesson
Homework: research 3 or interview 1
NEXT STEP — Scouting America's official Scoutly chatbot at scouting.org is built for this badge — use it (that's practicing Level 6). Then book your counselor session.
Imagine what it builds in the next ten years. The people steering it — deciding what it's used for, keeping it fair, keeping it honest —
artificial_intelligence.mb — merit badge complete_
🤖 generated with AI · verified by humans · powered by Scouts