Tools or Teammates? Examining Agency Negotiation in Human-GenAI Collaboration In Creative Work
Yuan Sun, Ignacio Cruz, Dong-Gyu Kim
Abstract
As Generative AI (GenAI) becomes increasingly integrated into creative work, creative professionals face an emerging dilemma: while they can leverage powerful AI capabilities through prompts, the automated process also introduces uncertainties and risks. This study, based on semi-structured interviews with 18 creative professionals, explores their experiences and perceptions of using GenAI tools in creative work. The findings revealed that GenAI’s perceived adaptability, creativity, and scalability enhanced sense of agency in the creative process. Conversely, limitations in customizability, uncertainty, and feedback mechanisms undermined human agency. These perceptions were influenced by users’ conceptualization of GenAI—as either a simple tool or a collaborative partner—and by the folk theories they developed through iterative communications with the technology. This research advances our understanding of human-machine collaborative dynamics in creative fields and provides insights for designing GenAI tools that better support creative professionals.