The principle of treating people and groups equitably in AI behavior, decisions, and outputs.
Fairness in AI is about reducing unjust or unequal outcomes across different people, groups, or perspectives. It matters whenever an AI system influences decisions, rankings, recommendations, or content that could reinforce bias.
In prompt engineering, fairness often involves choosing inclusive language, checking outputs for skewed assumptions, and evaluating whether the system behaves consistently across different inputs.
Fairness is closely connected to bias, but it focuses more directly on the quality of treatment and outcomes rather than only on the source of distortion.