Week 9 — What Is AI, Really?¶

At a glance


Date	Friday, October 16, 2026
Unit	2 — AI Unmasked
Big idea	AI is not magic or a mind. It is software that finds patterns in data and uses those patterns to make predictions.
Goal	Students leave with a plain-language definition of AI and a healthy question: "What data did it learn from?"
Student accounts needed	None. Quick Draw is no-login.
Prep time	5 minutes

Why this lesson looks the way it does¶

This is the reset after a one-week gap and the bridge from Unit 1. The students just spent eight sessions telling computers exactly what to do. Now you shift the question: what changes when a computer appears to make a judgment?

The key move is to demystify without deflating. A 5th grader can understand "pattern from examples." A 10th grader can push on the boundary between prediction, reasoning, and intelligence. Your industry perspective is valuable here, not because you need war stories, but because you can calmly separate hype from mechanics: in real organizations, AI is useful precisely because it is limited, measurable, and designed for a task.

Do not try to define every kind of AI. Give them one durable frame: data in, pattern learned, prediction out, human responsibility around it.

CSTA 2026 standards¶

Concept area	Connection
Algorithms & Design	Students compare rule-based programs with systems that classify by learned patterns.
Programming	Students connect Unit 1 instructions to the idea that AI systems are still software built by people.
Data & Analysis	Students identify training data as the source of an AI system's behavior.
Systems & Security	Students consider where an AI system runs and why school devices may restrict tools.
Computing & Society	Students discuss why AI claims should be examined rather than accepted as magic.

"I can…" objectives¶

By the end of today, students can say:

I can explain AI as software that learns patterns from examples.
I can tell the difference between a rule I wrote and a pattern a machine learned.
I can ask what training data an AI system used.
I can name one useful AI system and one way it might fail.

Teacher prep — the 5-minute version¶

Open https://quickdraw.withgoogle.com on the projector.
Optional: open a blank document or slide for a class T-chart.
Write on the board:

Last Friday: We had flex week. The Friday before: you shipped games.
Today: We unmask AI.
AI = examples → patterns → predictions
Under that, draw two columns:

Rules we write | Patterns it learns

Minute-by-minute run sheet¶

0:00–0:04 — Anchor¶

Start with the week gap on purpose.

"Last Friday we were not together because of the flex week. The Friday before that, you shipped games. In those games, nothing happened unless you gave the computer exact instructions. Today we start Unit 2, and the question changes: how can a computer seem to know something when we did not write every rule?"

Point back to Week 1:

"On the first day, Quick Draw guessed your drawings, and we ended with one sentence: It learned from us. This whole unit is us finally opening that sentence up."

0:04–0:12 — Mini-lesson: rules versus learned patterns¶

On the board, use a familiar example.

"If I write a MakeCode rule that says, 'if the pizza touches the player, add one point,' that is not AI. It is a rule. The computer follows it exactly.

If I show a system thousands of drawings labeled 'pizza' and thousands labeled 'not pizza,' and it learns what pizza drawings usually have in common, that is closer to AI. I did not write a pizza rule. I gave it examples."

Add the definition slowly:

"For this class, our working definition is: AI is software that uses data to learn patterns and make predictions or decisions."

Then add the professional lens:

"In business, the first serious question is almost never 'Is it intelligent?' The better question is, 'What job is it doing, what data did it learn from, and how will we know if it is wrong?'"

0:12–0:36 — Paired work: AI or not?¶

Use Driver/Navigator pairs. Students do not need a document; this can be verbal with scrap paper.

Give pairs this list on the board. They sort each item into AI, not AI, or depends:

A calculator adding 48 + 19
A game enemy moving left and right on a timer
Quick Draw guessing a doodle
A phone suggesting the next word in a text
A thermostat turning on at 70 degrees
A website recommending a video
A spell checker underlining a word
A camera app blurring the background

Say:

"Your job is not to guess what I think. Your job is to defend your category. If you say 'depends,' tell me what it depends on."

Circulate with questions:

"Where is the data?"
"Is it following a rule someone wrote, or using patterns from examples?"
"What prediction is it making?"
"How could it be wrong?"

Cross-age pairing move

Put older students in the Navigator role first. They must ask questions before giving answers. Younger students often have sharper intuition here because they are less trapped by technical vocabulary.

Ask for one easy case, one hard case, and one disagreement.

"Give me one item that is definitely not AI. Now give me one that probably is. Now give me one where reasonable people could argue."

Land the point:

"The label 'AI' is less important than the habit. When a system makes a guess, ask: What examples trained it? What is it predicting? What happens if it is wrong?"

0:42–0:45 — Teaser + prompt¶

"Next Friday you will teach a machine yourself. Not watch me do it — you will train it, test it, and see it make mistakes."

Tiered challenges¶

🟢 Starter (grades 5–6): Pick one example from the list and say what the computer is trying to guess.
🟡 Builder (grades 7–8): Explain why "AI" and "computer program" are not opposites. How can AI still be software?
🔴 Extender (grades 9–10): Choose a "depends" example and describe two possible implementations: one rule-based, one learned from data.

"What is one question we should ask before trusting an AI system?"

🤔 Think About This¶

This week, notice one place a computer sorts, recommends, filters, or guesses something for you.

Ask yourself: what examples might have trained it to make that guess?

🔌 Unplugged Plan B — if the internet is down¶

You do not need Quick Draw. Run the whole lesson with people as the classifier.

Time	Activity
0:00–0:04	Same anchor and Week 1 callback.
0:04–0:12	Mini-lesson with the board definition: examples → patterns → predictions.
0:12–0:25	Human classifier. Secretly choose a rule for "approved shapes" on index cards, such as "has a triangle." Show 8 examples marked yes/no. Students infer the pattern.
0:25–0:34	Groups create their own secret classifier rule and 6 example cards. Another group tries to infer the rule.
0:34–0:41	Debrief: examples can teach a pattern, but weak examples teach weak patterns.
0:41–0:45	Same teaser and Think About This.

Common problems & fixes¶

Problem	Fix
Students call every computer thing AI	Keep returning to: "Where are the examples? What prediction is being made?"
Older students argue definitions	Welcome it, then narrow: "For this unit, our classroom definition is the one on the board."
Quick Draw dominates the lesson	Limit it to the callback. Week 10 is the hands-on training day.
Students get fearful about AI	Reframe: useful tools can still fail. The goal is not fear; it is informed use.
A student asks if AI is alive	Answer plainly: "No. It can produce human-like outputs without being a living mind."

Vocabulary¶

Artificial intelligence (AI) — software that uses data to learn patterns and make predictions or decisions
Training data — examples used to teach an AI system
Prediction — a system's best guess based on patterns it learned
Rule-based program — software that follows instructions written directly by a person
Classifier — an AI system that sorts inputs into categories

Reflection — fill in after class¶

Which examples caused the best disagreement?
Who naturally asked about training data?
Did the Week 1 "It learned from us" callback land?
Pairing note for next week: who should be Driver first?

Next week: Week 10 — Training Data: Teach a Machine Yourself