VERSION B · Bac pro · B1 (lighter version)
Session 3: AI makes mistakes
Part 3. Can we believe everything AI says?
Objectives
- Understand fake news and deepfakes
- Check information
- Complain politely on the phone
Warm-up
A picture of the Pope in a white puffer jacket. Real or fake? It was made by AI in 2023.
Documents
Short text. In 2024, a chatbot on the Air Canada website gave a customer wrong information. A tribunal said the company had to pay the customer.
Words: fake news, deepfake, scam, phishing, fact-check, clickbait.
Activities
- Match the words and the pictures.
- Three questions before you share: Who? When? Other sources?
- Small task: phone call to complain about a chatbot.
Sentence starters
- Hello, I am calling because...
- Your chatbot told me...
- I would like...
- We are sorry for...
- We will...
Role play
Pairs: A is the customer, B works for the company. The chatbot gave a wrong price. A complains politely; B says sorry and offers a solution.
Gap-fill worksheet

Word bank: fake, check, scam, sorry, chatbot, deepfake
- This photo is ______.
- A ______ is a fake video made by AI.
- Always ______ the information.
- A ______ wants your money.
- The ______ made a mistake.
- The company said: we are ______.
Answer key
- fake
- deepfake
- check
- scam
- chatbot
- sorry
Homework
Write a short apology (80 words): what happened, sorry, a solution, a promise with will from the company to the customer.
Assessment grid
| Criteria | Level 1 | Level 2 | Level 3 | Level 4 |
|---|
| Phone call | Very short | Basic | Clear | Clear and polite |
| Apology | Missing | Sorry only | Sorry and solution | Complete and polite |
Links reminder