Ensure the AI system’s language and behaviors do not reinforce undesirable and unfair stereotypes and biases.
Enable the user to access an explanation of why the AI system behaved as it did.
Allow the user to globally customize what the AI system monitors and how it behaves.