Help the user understand how often the AI system may make mistakes.
Allow the user to globally customize what the AI system monitors and how it behaves.