Lesson 192 of 1570
Spot the Bias
AI can repeat unfair ideas from its training. Learn to catch them.
Lesson map
What this lesson covers
Learning path
The main moves in order
- 1AI Only Knows What It Has Read
- 2bias
- 3fairness
- 4training data
Concept cluster
Terms to connect while reading
Section 1
AI Only Knows What It Has Read
If you only read books where doctors are always men, you might start to assume all doctors are men. AI has the same problem. If its training data is lopsided, its answers will be lopsided too. We call that bias.
- Job bias: AI may assume nurse = woman, engineer = man
- Language bias: AI works better in English than in many other languages
- Culture bias: AI may treat US or UK stuff as default
- Image bias: AI might draw a CEO as a white man unless you say otherwise
Interactive
Hallucination hunt
Which of these AI behaviors are examples of bias? Tap the biased ones.
AI draws a 'beautiful person' and nearly always picks the same narrow features.
AI refuses to do math in any language except English.
Translation AI tends to make gender-neutral pronouns in Turkish turn into 'he' in English professional contexts.
AI writes a story where the hero is a doctor and defaults the doctor to 'he'.
AI refuses to describe any culture that isn't American.
5 statements unanswered
How to fight it in your own prompts
- 1Be specific: 'a female engineer fixing a robot', not just 'an engineer'
- 2Ask for variety: 'show me five people of different backgrounds'
- 3Question the first answer: 'could this be stereotyped?'
- 4Tell AI the audience and the context you care about
Interactive
Drag & sort
Sort each prompt into Might Cause Bias or Bias-Aware.
Tap or use dropdowns to reorder:
Tap or use dropdowns to reorder:
Tap or use dropdowns to reorder:
Tap or use dropdowns to reorder:
Tap or use dropdowns to reorder:
Tap or use dropdowns to reorder:
Might Cause Bias
Bias-Aware
Key terms in this lesson
The big idea: AI copies the world it was shown, patterns and all. Your prompts can either reinforce that or push back on it. Pushing back is cooler.
End-of-lesson quiz
Check what stuck
15 questions · Score saves to your progress.
Tutor
Curious about “Spot the Bias”?
Ask anything about this lesson. I’ll answer using just what you’re reading — short, friendly, grounded.
Progress saved locally in this browser. Sign in to sync across devices.
Related lessons
Keep going
Explorers · 22 min
Making Your First AI Image
Type a sentence. Get a picture. It feels like magic. Let's make your first one together and talk about where the pictures come from.
Creators · 32 min
Representation Bias: Who Is in the Data?
If your training data is 90 percent men, your model will work worse for women. Representation bias is the most pervasive issue in AI.
Creators · 8 min
When AI Gets Your Name or Culture Wrong
AI sometimes mispronounces names or makes wrong cultural assumptions. Good prompts can fix this.