JunoJR Learning
← AI Lab
Chapter 1 · Beginner · 45–60 min

Meet Your AI Sidekick

What AI really is, and how to talk to it.

In this chapter you'll

  • Explain in your own words what an AI chatbot is
  • Write and send your first clear prompts
  • Spot the moment an AI is guessing instead of knowing
  • Build a reusable AI helper that does one useful job for you

🛠️ You'll build: Your first AI helper — one saved prompt that does a real job, like turning your notes into a practice quiz.

Welcome, future AI engineer. 👋

Every big thing you'll build in this lab — chatbots, smart apps, even your own mini "agent" — starts with one skill: talking to an AI clearly. That's what this first chapter is about. By the end you'll have sent your first prompts and built a little helper you can actually use for homework, hobbies, or fun.

No code yet. Just you and your new sidekick.

What is an "AI sidekick", really?

The chatbots you'll use (like Claude or ChatGPT) are a kind of AI called a large language model, or LLM for short. Here's the whole idea in one sentence:

An LLM is a program that has read an enormous amount of text, and uses that to predict good words to write next.

That's it. When you type a question, it doesn't "look up" an answer like a search engine. It writes one, word by word, by guessing what a helpful response would look like — using patterns it learned from all that reading.

That one fact explains almost everything about AI:

  • It's amazing at language: explaining, rewriting, brainstorming, summarizing.
  • It sounds confident even when it's wrong, because it's predicting words that sound right (you'll dig into this in Chapter 4).
  • The clearer your instructions, the better its guess. That's the whole game, and it has a name: prompt engineering.

A prompt is just the message you send the AI. Engineering it means choosing your words on purpose to get a better result.

Before you start

Stay safe

AI chatbots are powerful, so a few ground rules keep you safe and in control:

  • Get set up with a parent or guardian. Most AI tools are for ages 13+ (some 18+, or 13+ with a parent's OK). If you're younger, do this lab on a grown-up's account, with them nearby.
  • Never share private information — your full name, address, school, phone, passwords, or photos. The AI doesn't need them, ever.
  • AI can be wrong or weird. Treat its answers like a confident classmate's: useful, but always worth checking.
  • Be kind and sensible. If anything an AI says feels off, upsetting, or unsafe, stop and tell a trusted adult.

You'll need: a device with internet, and an AI chatbot open in your browser (Claude or ChatGPT both work great). Ready? Let's talk to it.

Step 1 — Say hello and watch it think

  1. Open your AI assistant in a new tab.

  2. In the message box, type exactly this and send it:

    Hi! In one friendly sentence, explain what you are to a 12-year-old.

  3. Read the reply. Notice how it answers the way you asked — friendly, one sentence, for a 12-year-old. You shaped the output with your words. That's your first taste of prompt engineering.

Tip

See how the words appear one at a time? That's the model literally predicting the next word, then the next. It's not loading a finished answer — it's writing it live.

Step 2 — Try three kinds of asks

The same sidekick can do very different jobs depending on what you ask. Send these three prompts, one at a time, and notice how the type of answer changes:

  1. Explain something:

    Explain how a rainbow forms, in 3 short bullet points.

  2. Create something:

    Write a 4-line funny poem about a robot who is afraid of socks.

  3. Transform something (this one is a superpower):

    Rewrite this to sound excited and add emojis: "We are going to the museum."

You just used the three things LLMs are best at — explaining, creating, and transforming text. Almost every AI project is a mix of these.

Step 3 — Catch it guessing

Now let's see the model's weak spot, on purpose. Ask it something it can't really know:

What did I have for breakfast this morning?

It can't know that — it can't see you! A good assistant will say it doesn't know. But sometimes an AI will make something up confidently. That's called a hallucination, and spotting it is one of the most important skills you'll build.

Check yourself

Answer these in your own words (out loud or in a notebook):

  • An LLM mostly does what with words — looks them up, or predicts them?
  • What are the three things LLMs are great at?
  • What's a "hallucination", and why should you double-check important answers?

Project — Build your first AI helper 🛠️

Time to make something you'll actually reuse. A "helper" is just a really good prompt you save and use again whenever you need that job done.

You'll build a Notes → Quiz helper: paste in any notes, and it makes a practice quiz to test yourself.

  1. Start a fresh chat and send this prompt:

    You are my study buddy. I'll paste my notes. Make a 5-question practice
quiz from them: a mix of multiple-choice and short-answer. Don't show the
answers yet — put them at the very bottom under "Answer key".

  2. Then paste in some real notes (a few sentences from a class, a book, or a hobby). Send it.

  3. Read the quiz. Did it follow all your instructions — 5 questions, mixed types, answer key at the bottom?

  4. Improve it. Send a follow-up like: Make question 3 harder or Add one true/false question. Watch it adjust. Tweaking and re-asking is exactly how real prompt engineers work.

  5. Save your helper. Copy that prompt into a notes app or doc titled "My AI Helpers." You just shipped your first one. 🎉

Your turn

Make the helper your own. Change it into a different job, like:

  • A "explain like I'm 10" helper that simplifies anything you paste.
  • A story-starter that writes the first paragraph of an adventure using three words you give it.
  • A debate buddy that argues the opposite side of any opinion, nicely.

Keep the ones you like in your "My AI Helpers" doc.

Make it simpler · ages 9–11

Do it together with a grown-up. Pick the story-starter helper — it's the most fun. Give the AI three silly words (like "dragon, pizza, skateboard") and read the story out loud together. Then ask it to "make it sillier!"

Level up · ages 13–16

Push on why small wording changes matter. Run your quiz prompt twice: once ending with "keep it easy" and once with "make it challenging for a high-school student." Compare the results. Then add a constraint the model has to obey, like "every question must be answerable from the notes only." Notice how constraints make outputs more reliable — that idea powers everything from here on.

What you learned

  • An LLM predicts helpful words — it doesn't look up answers.
  • Your prompt shapes the output: explain, create, or transform.
  • AI can hallucinate, so you stay the fact-checker.
  • A saved prompt is a reusable helper — your first piece of AI "engineering".

You've earned the First Contact badge. 🏅

Next up: you'll learn the real techniques that turn an okay prompt into a great one — context, format, and examples. That's Chapter 2: Prompt Engineering Basics.

🏅 Finish this chapter to earn the First Contact badge.