Bias
Unfair or skewed patterns a model picks up from its training data or design and repeats in its outputs, like a hiring tool that favors one group of applicants.
“The screening model was rejecting older applicants, a bias it learned from past hiring.”
See also
In the history of AI
One of the Essentials
Bias is one of 20 Essential terms: the words you need to follow a news story or a work meeting about AI. It entered common use around 2023.