Help the user understand what the AI system is capable of doing.
Ensure the AI system’s language and behaviors do not reinforce undesirable and unfair stereotypes and biases.
Allow the user to globally customize what the AI system monitors and how it behaves.