AI›Level 1 · AI Explorer›Lesson 8

L1 · 08

Challenge: AI Detective Hunt

Time to put on your detective hat. Use everything from Unit A to crack a string of AI cases — is it really AI, and what did it learn from?

⏱ 1.5 hours🔥 Challenge lesson · no coding📚 After AI-L1-07🕵️ Detective worksheet
01

Learning Goals

5 min

By the end of this lesson you can:

  • Decide whether a real-world example is AI or not, with a reason.
  • Name the data an AI most likely learned from.
  • Spot a myth hiding inside a tricky case.
02

Warm-Up · The Detective's Question

8 min

We've covered a lot: what AI is, everyday AI, brains, myths, data and patterns. One question ties it together.

The golden question: "Did it learn from examples (data), or does it just follow a fixed rule?"

Quick warm-up case

A vending machine drops a drink when you pay. Not AI — fixed rule, no learning.

03

The Detective Toolkit

12 min

This is a challenge lesson, so just a quick recap of your three detective tools, then straight into the cases.

  1. Learn test: does it improve from examples, or always do the same thing?
  2. Data trace: if it's AI, what examples did it likely learn from?
  3. Myth check: is anyone claiming it has feelings, wishes, or perfect knowledge? That part is a myth.
Scoring

Give yourself 1 point for the right verdict and 1 for a good reason. Six cases — can you reach 12?

04

Worked Case · Solved Aloud

13 min

Case 0. Faridah's phone keyboard suggests "lemak" the moment she types "nasi".

  • Learn test: it learned which words usually follow which — that's learning. ✅ AI.
  • Data trace: huge amounts of typed text where "nasi" is often followed by "lemak".
  • Myth check: it does not "know" she is hungry — it predicts the next word. No myth claimed.

Verdict: AI, learned from text, no myth. Full marks. Now your turn.

05

The Cases · Crack Them All

20 min

For each case, write your verdict (AI / not AI) and a one-line reason. For AI cases, add the likely data. Reveal the answers only after you've tried.

  1. A traffic light changes on a fixed 60-second timer.
  2. A game boss dodges Imran's attacks better each time he plays.
  3. A photo app groups all the pictures of Kavitha together.
  4. A microwave beeps when its timer hits zero.
  5. A music app builds a new playlist it thinks Aaron will like.
  6. A "smart" toy claims it "loves you and feels lonely when you leave".
Reveal the detective's answers
  1. Not AI — fixed timer, no learning.
  2. AI — improves from examples of how Imran plays.
  3. AI — learned faces from many photos of Kavitha.
  4. Not AI — fixed rule: timer ends → beep.
  5. AI — learned from songs many people played together.
  6. AI + myth — it may recognise your voice (AI), but "loves you / feels lonely" is a myth — toys have no feelings.
06

Mini-Challenge · Build Your Own Case

12 min

Now flip the table. Write one tricky case of your own — something where it's not obvious if it's AI.

Swap with a partner and try to solve each other's case using the three tools.

It works if your case can be solved with a clear verdict and reason — not a trick with no answer.

Show a sample case

Case: "A weather app says there's an 80% chance of rain in Penang this afternoon." Answer: AI — it learned from years of weather data to predict; 80% is a confidence, not a promise.

07

Recap

5 min

A good AI detective asks one question: did it learn from data, or follow a fixed rule? Then they trace the data and check for myths. You've now finished Unit A — you can recognise AI in the wild.

Vocabulary Card

verdict
Your decision — here, "AI" or "not AI" — with a reason.
the golden question
"Does it learn from examples, or just follow a fixed rule?"
08

Homework · Three Real Cases

≤ 20 min

Find three real things around you. For each, write the verdict (AI / not AI), a reason, and — if it's AI — the data it likely learned from.