Reconfiguring Creative Labour: Generative AI, Work Practice, and the Ethics of Distributed Agency

The diffusion of generative artificial intelligence (GenAI) across the creative industries has sharpened debates about automation, productivity, authorship, and the future of work. Attention has so far gravitated towards two poles: macro-level questions about job displacement, and technical questions about model performance and output quality. Much less has been said about how these systems are folded into the daily routines of professional work, and how that folding reorganises the conditions under which creative labour is performed, judged, and legitimated. This paper examines the incorporation of generative AI into advertising agencies, treated as strategic observatories of wider transformations in cognitive and creative labour. Drawing on a multi-method qualitative design that combines a scoping review with thirty-three in-depth semi-structured interviews with creative professionals, it traces how AI-assisted workflows are reorganised around selective delegation, cognitive offloading, iterative evaluation, orchestration, and what it terms curatorial labour: the emergent expertise of directing, evaluating, and taking responsibility for what algorithmic systems generate. Drawing on socio-technical accounts of distributed creativity, the paper argues that generative AI redistributes rather than replaces creative production, dispersing expertise, evaluative authority, and aesthetic judgment across human actors, machines, datasets, and commissioning organisations. It further argues that AI governance remains too model-centred and proposes a workflow-centred perspective on generative AI at work.