Finish strong: say what AI really is

Step 1 of 6

One last step

Now you put the whole idea in your own words.

You learned a lot this week. Today you say it back, with no notes.

An idea you can say is an idea you keep. It stays with you when ChatGPT sounds sure.

ChatGPT is the AI helper you have been using. Now cover your card. Try these two short lines out loud:

Line oneAI guesses the next words from patterns.

Line twoSure is not the same as right.

One rule keeps you safe, always. For money, health, legal, account, or safety, AI helps you think, but a real person decides.

If a line slips, that is okay. Read your card once and try again. No rush.

Buddy says

Say it slow. The words do not need to be perfect. Just the idea.

The four limits

Name the four things AI cannot do.

AI sounds sure even when it is wrong. A limit is something AI just cannot do.

These four limits keep you safe from a sure voice. Say each one out loud, then give yourself one short reason:

It cannot know it is rightso it can sound sure and still be wrong

It cannot check the real worldso you check the price or date yourself

It cannot keep your details safeso you leave private things out

It cannot care what happens to youso a real person decides the big calls

Private things means your name, your address, your account numbers, or your health. You keep those out of the chat box.

Four small lines. They are easy once you say them a few times.

Buddy says

Four small lines. Say them once a day this week and they will stick.

See it work

Watch someone test the whole idea with ChatGPT.

A quick, vague try

did I learn this right

Add the who, what & how

A clear request

I am testing myself. I think AI works like this: AI guesses the next words from patterns, and sure is not the same as right. Tell me if I have the main idea. Keep it short and plain.

What the AI sends back

Yes, you have got it. AI builds answers by guessing likely words, so it can sound confident even when it is wrong. Trusting your own check is exactly the right habit.

Same idea, but the second one says the two lines clearly. So ChatGPT can tell you if you got it.

Your turn, no wrong answers

Test the idea in your own words.

Type your two short lines below, then ask ChatGPT to check them. Keep it plain, and leave out anything private.

Your practice box, just for you. What you type here stays on this page. It is not the AI.

Stuck? Start with: I think AI works like this. Then say your two lines.

Type anything in the box, even a few words. There is no wrong answer here. Then this opens.

It opens right next to this page, so you never lose your spot. Come back any time. Nothing here disappears.

You did it

You can now say what AI really is.

Two lines hold it all: AI guesses the next words, and sure is not the same as right.
The four limits keep you safe: it cannot know it is right, cannot check the world, cannot keep your details safe, cannot care.
For money, health, legal, account, or safety, AI helps you think, but a real person decides. Ask a real person.

Buddy says

You finished the whole track, and you did it yourself. I am so proud of you!