Abstract
Striking the appropriate balance between humans and co-creative AI is an openresearch question in computational creativity. Co-creativity, a form of hybridintelligence where both humans and AI take action proactively, is a processthat leads to shared creative artifacts and ideas. Achieving a balanced dynamicin co-creativity requires characterizing control and identifying strategies todistribute control between humans and AI. We define control as the power todetermine, initiate, and direct the process of co-creation. Informed by asystematic literature review of 172 full-length papers, we introduce MOSAAIC(Managing Optimization towards Shared Autonomy, Authority, and Initiative inCo-creation), a novel framework for characterizing and balancing control inco-creation. MOSAAIC identifies three key dimensions of control: autonomy,initiative, and authority. We supplement our framework with controloptimization strategies in co-creation. To demonstrate MOSAAIC's applicability,we analyze the distribution of control in six existing co-creative AI casestudies and present the implications of using this framework.