Bias refers to systematic errors in AI results that arise when training data or models favor or disadvantage certain groups or perspectives. For example, a hiring filter trained on historical data might disadvantage women because more men were hired in the past. Bias mitigation includes diverse data, audits, and fair training methods.
In short
What is bias in AI?
Bias is a systematic distortion in AI results, often caused by one-sided training data. A hiring filter might disadvantage women, for example, if trained on historically male-dominated data. Countermeasures include diverse data, audits, and fair training methods.
Using AI in your organisation?
We show you which applications are worth it for your organisation.