Enable the user to access an explanation of why the AI system behaved as it did.
Limit disruptive changes when updating and adapting the AI system’s behaviors.
Immediately update or convey how user actions will impact future behaviors of the AI system.