Human-AI Interaction Patterns

Summary: To improve human trust in AI systems, designers need to provide responsible and instant feedback to users while they use the AI features in mobile applications to help them understand the capabilities and limitations of AI systems. In this work, we analysed the existing practices and summarise a taxonomy of interaction patterns to operationalise high-level responsible UI design principles and guide the designers to design more reliable and human-centred interactions for AI-powered apps.

Type of pattern: Product pattern

Type of objective: Trustworthiness

Target users: Designers

Impacted stakeholders: Development teams, AI users, AI consumers

Relevant AI principles: Human, human-centred values, reliability and safety, transparency and explainability

Context:  With the emergence of deep learning techniques, mobile apps are now embedded AI features for enabling advanced tasks like speech translation, to attract users and increase market competitiveness. A good interaction design is important to make an AI feature usable and understandable, and gradually build the user trust in the underlying AI systems. However, AI features have their unique challenges like sensitiveness to the input, dynamic behaviours and output uncertainty. Existing guidelines and tools either do not cover AI features or consider mobile apps which are confirmed by our informal interview with professional designers. While some provides high-level design principles, it is hard for designers to interpret and implement these guidelines.

Problem:  What kinds of AI Features are supported in mobile apps? What interaction patterns are used in mobile AI apps?

Solutions: We collected a large-scale mobile applications and identified AI-powered apps. By adopting an open-coding strategy, we cultivate the first taxonomy that summarises the existing interaction patterns for AI features. Our work could facilitate and guide the designers, who do not have specific knowledge in AI, to design reliable and understandable interaction designs for AI-powered apps.


  • Increased trust: providing end-users with instant feedback about the capabilities and limitations of AI features can increase their trust when using these features.


  • Assessment: We only provide a comprehensive taxonomy of user-AI interaction patterns to designers, without criticising the quality, advantages and  disadvantages of each pattern.

Related patterns:

  • AI mode switcher:  This work could work with AI mode switcher to give users freedom to decide whether use AI features or not.

Known uses:

  • Google is developing PAIR to provide high-level design patterns for AI systems.
  • Microsoft also provide Human-AI Interaction Guidelines to recommend the best practices of how AI systems should interact with human