Aim:
To understand how bias can affect AI results and explore ways to make AI fairer.
Materials:
-
Printed pictures or face cards of people with diverse ages, genders, and appearances
-
Computer with internet access (optional)
-
Google Teachable Machine (optional for demo)
-
Chart paper and sketch pens
Steps:
Step 1: Discussion – What is Bias?
Ask students:
-
What does bias mean in real life?
-
Can you think of a situation where someone was treated unfairly?
Relate it to AI:
“AI can also become unfair if it learns from limited or biased data.”
Step 2: Demonstration (Optional)
If computers are available, open Teachable Machine and create a quick image model with limited diversity:
-
Train only on a few types of faces or gestures.
-
Test it with different faces (varied gender, skin tone, glasses, etc.).
Discuss what happened:
Did it recognize everyone correctly? If not, why?
Observation: The AI made mistakes because it didn’t have enough varied examples.
Step 3: Reflection Discussion
Ask students:
-
How can we fix the AI’s mistakes?
-
What kind of data should we collect next time?
Emphasize:
Adding diverse and balanced data improves fairness and accuracy.
Step 4: Creative Task – Make a Poster
Title: “How to Make AI Fair”
Students draw or write ideas like:
-
“Use data from everyone.”
-
“Check results before launching.”
-
“Include people of all kinds in AI design.”
Posters can be displayed in the classroom or STEM corner.
Reflection
Ask students to think and write:
-
What does fairness mean to me?
-
How can students or engineers make technology more inclusive?
Key takeaway:
“AI is only as fair as the people who build and train it.”