For high school · About 15 minutes
AI in real life
Choose an activity to begin. After the four activities, try the final challenge.
Activity 1 of 4 3 min
All activitiesCheck the receipt
Predict
Would you pay €24 or €28? Read what the chatbot says, then look for the clue on the receipt.
Try
Select the line that settles it.
Hoodie receipt
- Price before VAT
- €20
- VAT included
- €4
- Total paid
- €24
Compare with a correction
You paid €24. The receipt says the €4 VAT is included, so adding it again counts it twice.
Try a different receipt
A receipt says €30 total, including €5 VAT. What did you pay? Explain why in one sentence.
See the answer and 20-second walkthrough
You paid €30. The €5 VAT is already part of the total.
Read the transcript
A chatbot reads a receipt and says you owe thirty-five euros. Let’s check. The receipt lists five euros of included VAT. That word, included, matters. The total says thirty euros. That’s the amount you pay. The five euros are already in the total. Adding them again counts the tax twice.
Explain
In one sentence, explain which word tells you whether to add the VAT again.
Activity 2 of 4 4 min
All activitiesTrain a filter
Predict
Will “claim class reminder” be marked spam, regular, or unsure? Guess before you test it.
Try
The filter gives each distinct word one point if it appears in just one group of training examples. Both groups get zero for shared or unseen words. A tie means unsure.
Training examples
- Spam “claim prize”, “claim reward”
- Regular “class update”, “class reminder”
Test messages
Pick a message. Then add the extra spam example to see the result for the same message again.
Explain
Which words stop earning points after you add the example? Try all three messages and explain why one prediction stays the same.
Compare with an explanation
“Class” and “reminder” now appear in both groups, so they stop scoring. “Claim class reminder” goes from regular to spam, and “class reminder tonight” goes from regular to unsure. “Claim prize tonight” stays spam because “claim” and “prize” still earn spam points, and the filter has never seen “tonight.”
Activity 3 of 4 2 min
All activitiesChoose a safe prompt
Predict
You’re studying how to read a paycheck. Your teacher allows AI for practice. What information would you keep out of the prompt?
Try
Pick the prompt that helps you practice safely.
Explain
How would you check an AI-generated practice answer yourself?
Compare with an explanation
Solve the practice question using the fictional pay stub, then compare your result with the answer. Follow your teacher’s rules for AI use and disclosure on graded work.
Activity 4 of 4 3 min
All activitiesAudit AI answers
Predict
Which kind of document might make a chatbot guess?
Try
The chatbot got the same instruction for each document.
“Copy the tax amount shown. If no tax amount is shown, say ‘not shown.’”
Explain
How many replies match their documents? What should the chatbot say when the information is missing?
Compare with the answer key
Three of four replies match. The price tag shows a $20 price but gives no tax amount. The chatbot should say “not shown.”
3 min
All activitiesFinal challenge
Check the chatbot’s answer against this fictional pay stub.
Pay stub
- Gross pay:
- $120
- Federal income tax withheld:
- $7
- Other deductions:
- $10
- Take-home pay:
- $103
Type your answer or say it aloud before you reveal the model answer. Anything you type stays on this page.
Reveal the model answer
Correction. Federal income tax withheld was $7. The $17 difference also includes $10 in other deductions.
Evidence. “Federal income tax withheld: $7.”
For teachers
Students can work through the 15-minute lesson on their own. In pairs, have one student predict and the other explain, then swap roles. Use the final challenge as an exit ticket and look for both a correction and an exact line of evidence.