Fairness in AI
Fairness in AI is the goal that AI systems treat people equitably and do not discriminate unjustly against individuals or groups.
What is Fairness in AI?
There are several mathematical definitions of fairness, such as equal error rates or equal selection rates, and they cannot all be satisfied at once, so choosing one is a value judgement.
How it works
Teams define fairness metrics, measure them across groups and adjust data, models or decision thresholds.
← All Artificial Intelligence & Superintelligence definitions · ← Back to the A–Z definitions