AI Bias
When a model's outputs unfairly favor or disadvantage certain groups, usually because its training data carried those patterns.
It is a real risk anywhere AI touches people, from hiring to lending, and a reason outputs need human review.
Frequently asked questions
Where does AI bias come from?
Usually from the data the model learned on. If the training text reflected unfair patterns in the real world, the model can absorb and repeat them, so the bias is inherited from history rather than invented by the machine.
Why is AI bias a serious concern for businesses?
Because anywhere AI touches people, like hiring, lending, or evaluating customers, biased outputs can unfairly favor or disadvantage certain groups. That creates real fairness and legal risk, which is a big reason such outputs need human review.
Can AI bias be fixed completely?
It can be reduced through careful training, testing, and guardrails, but it is hard to eliminate entirely because it reflects patterns in messy real-world data. The practical answer is to watch for it and keep a person reviewing decisions that affect people.
New to all this? Start with what an AI agent really is, browse the full glossary, or explore the learning hub.
← Back to the glossary