Proposal of a Facilitation and Process Model for Enhancing Creativity in Co-design Workshops with Generative AI: The Use of ChatGPT

Abstract

Background In today’s complex and multidimensional societal issues, interdisciplinary collaboration is essential, and cooperative and collective ‘creativity’ is increasingly highlighted as a crucial factor in addressing such challenges. While endeavors to enhance creativity through artificial intelligence (AI) collaboration persist across diverse sectors of society, literature addressing creativity within the framework of co-design and studies aimed at maximizing creativity are scarce. Therefore, this study aims to propose an effective facilitation role utilizing generative AI in the co-design process that enhances participants’ creativity and to suggest a workshop process model for co-design utilizing generative AI. Methods Through theoretical considerations, principles for executing co-design were defined and creativity elements and facilitation roles were derived based on these principles. Building upon these concepts, a workshop process and toolkit for fostering creativity through collaboration between humans and generative AI were designed. Workshop experiments were conducted, employing different treatments regarding the provision of ChatGPT (Chat Generative Pre-trained Transformer), to compare and verify the effectiveness of two workshop sessions with the same group. Observational analyses of the workshop process, reflective interviews with participants, and expert evaluations of the outputs were conducted to derive insights. Reflecting on these findings, a blueprint for a workshop process model leveraging generative AI for enhancing creativity was proposed. Results 10 principles of co-design execution were derived, and 5 creativity elements were identified along with 15 sub-elements in accordance with these principles. Corresponding to the sub-elements of creativity, 14 elements of creativity facilitation were identified. The utilization of ChatGPT positively influenced various aspects of creativity in co-design, such as facilitating collaborative knowledge [CC1], fostering collaborative design thinking [CC3], aiding idea expression [CC4], and supporting the collaborative environment [CC5]. As a supplementary facilitator, ChatGPT played a useful role in providing knowledge support, aiding idea expression, and facilitating collaboration. Comparing the evaluation results of creative output, while usefulness remained consistent, fluency was higher in ChatGPT workshops. Conversely, originality was rated higher in conventional workshops. This suggests that while the utilization of ChatGPT contributed to increasing productivity in workshops, ChatGPT may have limitations in enhancing the quality of ideas. Conclusions The collaborative role of AI as a facilitator for enhancing creativity generally had a positive impact on participants’ creativity. However, depending on participants’ intrinsic motivations, AI could potentially act as a hindering factor to creativity. Thus, continuous intervention by human facilitators is deemed crucial to support the evolution of high-quality ideas. This study validates the utility of utilizing generative AI as a collaborative creativity facilitation tool through empirical research in codesign workshops and suggests the direction for effective collaboration between human and generative AI based on the co-design process. From a practical standpoint, proposing a concrete blueprint for a codesign workshop process model utilizing generative AI is vital for addressing practical problem-solving in complex and interdisciplinary co-design activities. This model can be applied across various problem-solving domains in the future.