Bias in AI
How training data can skew AI responses.
AI learns from data created by people — and that data can reflect stereotypes, gaps, or imbalances. The model may favour certain viewpoints, languages, or demographics.
Be aware of bias when AI gives advice about people, hiring, health, or culture. Use diverse sources and human judgment.
Bias in practice
Situation
A hiring manager uses AI to screen CV summaries. The model favours certain university names because those patterns appeared often in its training data.
Outcome
Human review catches the pattern. The team adds structured criteria and never relies on AI alone for hiring decisions.
Watch out for
✕ Treating AI advice about people as neutral
Instead: Consider bias and seek diverse perspectives.
✕ One-size-fits-all health or cultural advice
Instead: Consult qualified professionals for personal decisions.
Knowledge check
Why can AI responses be biased?
Choose the best answer, then check your understanding.