Interactive Journaling with AI: Probing into Words and Language as Interaction Design Materials

Abstract

Conversational Agents (CAs) are making human-computer interaction more collaborative and conversational through using natural language. The HCI and interaction design communities, have been experimenting with and exploring the area of designing conversational interactions. Furthermore, interaction designers may need to acquire new skills for designing, prototyping, and evaluating artifacts that embody AI technologies in general, and CAs in particular. This paper builds upon a previous study on principles of designing interactive journaling experiences with CA and explores the practice of designing such experiences, using words, language, and conversations as design materials. We present a prototype for interactive and reflective journaling interaction with CA and the result of a Wizard of Oz experiment. Our findings suggest that designing interactions with CA challenges designers to use materials with inherently different natures and qualities. Despite this challenge, words appear to have unique characteristics to support designers to externalize and iterate on ideas, e.g., tone and intent. Hence, we suggest considering words, language, and conversations as the primary design materials, and the AI’s predictability, adaptivity, and agency as secondary materials, while designing human interactions with Conversational Agents.