Fragile by Design? Examining Organizational Resilience as Generative AI Absorbs HR Functions
Year: 2026
Type: article
Abstract: Generative AI now produces many of the records through which organizations enforce accountability, document compliance, and sustain institutional memory. Research on GenAI and knowledge management has largely treated this as a gain in efficiency and quality. We argue that it conceals a loss. Producing records is also how practitioners develop the capacity to evaluate and recover them. When GenAI takes over production, records continue to look sound, so their quality no longer reveals whether that capacity still exists. Organizations therefore become fragile without noticing, and discover it only when a record is contested or must be reconstructed. Building on IS infrastructure studies and research on high-reliability organizing, we develop a framework that relates how deeply AI has absorbed record production to how much human capacity is retained, yielding four resilience configurations: robust, brittle, adaptive, and blind. We ground the framework in two illustrative vignettes from HR functions, where record production is dense, accountability-laden, and observable, and outline a case study program to examine it.
Source: AIS Electronic Library (AISeL)
Author: Wu, Hongyu
Institution: Universitat de Barcelona
Language: English
Subfield: Management Information Systems
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7225762608
