ML concept that ensures decisions remain fair by being unaffected by sensitive attributes, such as race or gender, in hypothetical scenarios where these attributes are altered.
Generality: 590
