The emergence of Generative AI (GAI) has led to the development of a novel intent-based interaction paradigm for human-AI engagement. While plenty of works have developed guidelines for human-AI interaction over the past two decades, many of these may not be suitable for the design of GAI applications. In this paper, we propose 11 design guidelines in a structured framework that outlines key moments in human-GAI interaction. We validated the guidelines through multiple rounds of evaluation, including a design workshop where 33 user experience (UX) practitioners utilized the proposed guidelines to design and evaluate five GAI-driven products. The results verify the relevance of the guidelines to GAI design and the practical value of the framework. In addition, we discuss insights gained during the application of the design guidelines, as well as existing gaps in current GAI design knowledge. We argue that the proposed guidelines can serve as a practical resource for practitioners designing applications based on generative technologies and as a foundation for researchers exploring further design patterns in human-GAI interaction.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Design Guidelines for Human-Generative AI Interaction

  • Li Lin,
  • Yu Wang,
  • Yayu Ping,
  • Jian Gao,
  • Zongbo Wang,
  • Shouyu Wang

摘要

The emergence of Generative AI (GAI) has led to the development of a novel intent-based interaction paradigm for human-AI engagement. While plenty of works have developed guidelines for human-AI interaction over the past two decades, many of these may not be suitable for the design of GAI applications. In this paper, we propose 11 design guidelines in a structured framework that outlines key moments in human-GAI interaction. We validated the guidelines through multiple rounds of evaluation, including a design workshop where 33 user experience (UX) practitioners utilized the proposed guidelines to design and evaluate five GAI-driven products. The results verify the relevance of the guidelines to GAI design and the practical value of the framework. In addition, we discuss insights gained during the application of the design guidelines, as well as existing gaps in current GAI design knowledge. We argue that the proposed guidelines can serve as a practical resource for practitioners designing applications based on generative technologies and as a foundation for researchers exploring further design patterns in human-GAI interaction.