Understanding and influencing decisions

Show capabilities and limitations

Before using an AI assistant, a person sees information about what it can and cannot do.

Description

Explain to people what a feature or service can and cannot do both at the beginning, and in context, throughout their experience. Using short, authoritative statements can help users to give or remove their trust as they see appropriate.

For example, before first using a chatbot, show statements that explain the most important limitations a person may encounter. Later, remind them in context as any relevant limitations are reached during their journey.

IF thinks this pattern is important because it is difficult for users, and other people who may come into contact with the system to understand what different AI models can and cannot do.

Advantages

  • Helps people calibrate their trust
  • Helps people use AI successfully because they understand capabilities and limitations
  • Sets realistic expectations that AI is not infallible

Limitations

  • May reduce confidence in the ability of an AI feature if not implemented well
  • Providing just the right amount of information requires research
  • Using the wrong tone of voice can lower perceived credibility

Examples

  • Youtube →

    Shows parents what they can and can't see when setting up supervision features for teens

  • Grammarly →

    Shows contextual capabilities and limitations as people move through a choice funnel

  • Claude →

    Explains what its models excel at and how safeguarding works during onboarding

Related patterns