Is AI taking generativity further? An explorative study using the mirroring hypothesis

This paper explores how Generative Artificial Intelligence (GenAI) reshapes organizational architectures and creative work through the lens of the mirroring hypothesis. While GenAI systems are often celebrated for their generative capacities, their actual impact on innovation and organizational structures remains underexplored. Drawing on the mirroring hypothesis, which posits a structural isomorphism between technologies, tasks, and knowledge, we propose a conceptual framework mapping how di erent types of GenAI (agnostic, holistic, and specialized) reconfigure organizational architectures. We identify key mechanisms through which GenAI augments, excludes, ports, or substitutes both tasks and knowledge. This conceptual development is supported by an empirical illustration in the healthcare industry, where GenAI was used to accelerate mobile app code migration. The case reveals not only productivity gains but also a deep reconfiguration of developer roles and organizational routines. We formulate hypotheses on how GenAI's architectural nature and the characteristics of organizational knowledge influence task reorganization. Ultimately, we argue that GenAI's generativity lies less in its computational novelty and more in its ability to enable new organizational combinations. This work contributes to understanding the co-evolution of AI systems and organizational design, and extends the mirroring hypothesis to the context of generative technologies.