Co-creation With and For AI: The Case of Mental Well-Being AI Companion

Paper in proceeding, 2024

This study explores the collaborative design process involving humans and generative AI in designing an AI companion for knowledge workers, which consists of an ecosystem of different tools and interfaces for the knowledge worker, family and friends, colleagues and managers, as well as occupational and public health support. The research involves various co-design and human-centered design methods, such as stakeholder workshops, co-creation workshops, and diary studies. Generative AI was effectively used throughout the process for generating co-design probes and tools, analyzing qualitative data, and synthesizing the results and ideas into a final concept. Some of the main themes revealed by the qualitative synthesis of data include dual functionality, user control and data privacy, adaptive support, interfacing and trusted networks, and balancing health and work-life. The study also shows a tendency toward using AI companions not only for mental well-being but also for easing work tasks, indicating a blurring of personal and professional domains, a trend less noted in existing literature. The use of generative AI through a co-design process shows its potential for facilitating the process in different ways, such as summarizing information, data analysis and synthesis, unifying knowledge when participants are from different backgrounds, and synthesizing diverse ideas into final concepts. It also faces challenges, including variability in outputs and the time-consuming process of integrating various elements into prompts.