MI-CCy Quantifier: A Framework for Quantifying Mixed-Initiative Co-creativity in Human-AI Collaborations

Solange Margarido, Licínio Roque, Penousal Machado, Pedro Martins

Abstract: The study of creative collaboration between humans and computational systems is a recurring topic in the area of computational creativity. With the rapid developments of Generative AI systems with increasingly improved performance and greater agency, this topic assumes growing importance. In this paper, we explore a specific paradigm of human-AI collaboration where the participants in the creative process work as colleagues – mixed-initiative co-creativity (MI-CCy). We investigate parameters that can serve as guidelines for the implementation of AI systems that have both a high level of mixed-initiative and a high level of co-creativity. To this end, we introduce the MI-CCy Quantifier, a framework that researchers and practitioners can use for assessing co-creative tools on their degree of MI-CCy through scales of gradable criteria presented in an intuitive visual layout. We reflect on the applicability of the MI-CCy Quantifier by demonstrating its usage in the analysis of a co-creative system, and we conclude by proposing future directions for mutually beneficial collaborations between humans and AI.