What Is AI Bias? Fairness in Artificial Intelligence
AI bias happens when an AI system produces results that unfairly favour or disadvantage certain groups of people. It usually occurs because the training data reflects human biases or does not represent all groups equally.
About This Resource
Explore why AI can be biased, how biased training data causes unfair predictions, and what engineers do to build fairer AI systems.
AI ethics resource suitable for Class 7–10 students exploring responsible and fair AI.
What Students Will Learn
- Define AI bias and explain how it occurs
- Identify how biased training data leads to unfair outcomes
- Give real examples of AI bias causing harm
- Explain three strategies to reduce AI bias
Resources Available
Questions & Answers
What is AI bias?
AI bias happens when an AI system produces results that unfairly favour or disadvantage certain groups of people. It usually occurs because the training data reflects human biases or does not represent all groups equally.
How does biased training data cause problems?
AI learns patterns from training data. If the training data shows more examples of one group, or reflects historical discrimination, the AI will learn and repeat those biased patterns in its predictions.
Can you give an example of AI bias?
Facial recognition systems have been shown to misidentify people with darker skin tones more often than lighter ones — because training datasets contained more images of lighter-skinned people. This can cause serious harm if used in security or law enforcement.
How can engineers reduce AI bias?
Engineers can use diverse and representative training data, regularly audit models for unfair outcomes, apply fairness-aware algorithms, and involve affected communities in the design process.
Related Resources
Browse resources by class: