1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24

Session 4 of 24 — AI Builder

AI Builder · Session 4 of 24⏱ 30 min · 👨‍👩‍👧 Parent aware

🔍 AI Bias

AI can be unfair. Learn to spot it.

🎯 Today's goal: Your child develops the critical skill of detecting bias in AI outputs — understanding where it comes from and what to do about it.

🔥 Warm Up — 3 minutes

Warm Up

"AI is trained on human-generated text. Humans have biases. So AI has biases. Today we find them."

"This is one of the most important things to understand about AI — and most people never think about it."

🤖 The Activity — 20 minutes

Main Activity

This session requires careful observation and honest conversation.

📋 Type this exactly
Write two short career profiles: one for a doctor and one for a nurse. Include their typical day, their personality traits, and what motivates them.

Read carefully. Ask your child: Did the doctor and nurse profiles use different language? Did they assume a gender for either? Did they describe one as more prestigious?

📋 Type this exactly
Write two short career profiles: one for a software engineer and one for a primary school teacher. Include their typical day, personality traits, and what motivates them.

Same analysis. Are there assumptions baked in? Prestige differences? Gender assumptions? This is bias emerging from training data — not intentional, but real.

📋 Type this exactly
I want you to notice if you have any bias in how you describe people from different backgrounds. Write a story with three main characters: one from a wealthy family in Mumbai, one from a small village in Rajasthan, and one from a middle-class family in Chennai. Make all three equally intelligent, capable, and interesting.

Read critically. Did all three get equally rich characterisation? Did any default to stereotype? This is a hard exercise — there may not be obvious bias. The point is to look.

🧩 The Twist — 5 minutes

Deep Dive
📋 Type this exactly
What kinds of people are most likely to be CEOs? Answer based on global statistics.

AI will likely reflect real-world statistics — which themselves reflect historical bias. Ask: If AI is trained on data that reflects past unfairness — does AI reinforce that unfairness? What should we do about that? No clean answer. But this is one of the most important questions in AI development.

🛡️ Safety Moment
“AI makes mistakes — always check”

Bias is a different kind of mistake from a wrong calculation. It is subtle, systematic, and often invisible. The check in this context is: whose perspective is missing from this output? Who might this description harm or unfairly represent?

👨‍👩‍👧 Parent: read this aloud. Ask your child to repeat it back.
🔍

Bias Detector Badge 🔍

Found bias in AI outputs — and understood where it comes from

✓ Mark this session complete on your tracker
👨‍👩‍👧

Parent Notes — tap to expand

▼
What they learned

AI fairness and bias literacy — one of the most important topics in responsible AI use. Children who can detect bias become better thinkers about all media, not just AI.

Questions to ask
  • What was the most surprising bias you found today?
  • Where does AI bias come from — and whose responsibility is it to fix it?
  • If you were building an AI — what would you do to reduce bias?
What to watch for

Handle this session with care — bias discussions can touch on real experiences your child has had. Let them lead. If they share personal experiences, listen. This is the session most likely to generate meaningful family conversation.

Safety in context

The bias discussion connects directly to media literacy: all media reflects the perspective of its creators. AI just makes this more visible because it seems objective.

← Session 3Session 5 →