Help the user understand how often the AI system may make mistakes.
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.