When AI creates: Creative self-efficacy as a mediator in design professionals' adoption of AI image generators

Generative AI has introduced tools that do not merely support human work but produce creative outputs, raising questions about what drives their adoption among creative professionals. Unlike conventional productivity tools, AI image generators implicate users' perceived creative capability, not only their technical competence. This study integrates the Unified Theory of Acceptance and Use of Technology (UTAUT) and creative self-efficacy theory to examine whether creative self-efficacy (CSE) functions as a psychological mechanism linking technology acceptance factors to adoption intentions and use among creative professionals. Using survey data from 1011 design professionals, the study applies partial least squares structural equation modeling to examine adoption intentions toward AI image generators. The results show that CSE mediates the effects of three of the four UTAUT constructs on behavioural intention. Effort expectancy, however, has no direct effect on behavioural intention and operates only through CSE. For designers, this suggests that adoption is shaped less by whether they believe they can use the tool than whether using it affects how they perceive their own creative capability. Moderation analyses further show that these mediating pathways vary by gender and experience with AI, indicating that the relationship between perceived AI adoption factors and designers' creative self-perceptions differs across subgroups. The findings suggest that technology acceptance constructs may operate differently in creative domains, and that adoption models should account for creative self-efficacy when the technology itself performs creative tasks. Implications for theory and the adoption of generative AI in creative practice are discussed.