Help the user understand what the AI system is capable of doing.
Help the user understand how often the AI system may make mistakes.
Time when to act or interrupt based on the user’s current task and environment.
Display information relevant to the user’s current task and environment.
Ensure the experience is delivered in a way that users would expect, given their social and cultural context.
Ensure the AI system’s language and behaviors do not reinforce undesirable and unfair stereotypes and biases.
Make it easy to invoke or request the AI system’s services when needed.
Make it easy to dismiss or ignore undesired AI system services.
Make it easy to edit, refine, or recover when the AI system is wrong.
Engage in disambiguation or gracefully degrade the AI system’s services when uncertain about a user’s goals.
Enable the user to access an explanation of why the AI system behaved as it did.
Maintain short-term memory and allow the user to make efficient references to that memory.
Personalize the user’s experience by learning from their actions over time.
Limit disruptive changes when updating and adapting the AI system’s behaviors.
Enable the user to provide feedback indicating their preferences during regular interaction with the AI system.
Immediately update or convey how user actions will impact future behaviors of the AI system.
Allow the user to globally customize what the AI system monitors and how it behaves.
Inform the user when the AI system adds or updates its capabilities.
