This refers to AI unintentionally generating inappropriate information by reflecting prejudices and social biases learned from training data.

Return to glossary list