Governing Artificial Intelligence as a Macro Socio-Technical Regime: The F-I-C-P Framework for Regime-Level AI Infrastructure and Governance

Abstract

Artificial intelligence (AI) is often analysed through technical, organisational, or regulatory perspectives that treat capability, adoption, and governance as separate issues. This article conceptualises AI instead as an emerging macro socio-technical regime generated through the interaction of finance, infrastructure, capability, and productivity. It develops the F-I-C-P framework to explain how finance enables infrastructure, infrastructure conditions capability, and capability creates the possibility--but not the guarantee--of productivity and value realisation. The article also introduces the concept of regime-level AI infrastructure to identify concrete configurations, such as AI Gigafactories, foundation-model platforms, national compute infrastructures, public AI platforms, sectoral AI ecosystems, and corporate AI infrastructures, through which the macro-regime becomes observable and evaluable. The central analytical problem is structural misalignment: the risk that financial expectations, infrastructural expansion, technical capability, organisational absorption, productivity realisation, governance capacity, and societal outcomes evolve at different speeds or under unequal conditions of power. Governance is therefore conceptualised not as a final stage after technological deployment, but as a transversal and anticipatory steering capacity. Briefly illustrated through Spain and Chile as non-frontier configurations, the article argues that AI governance should be understood as regime alignment: the task of anticipating, evaluating, and correcting structural misalignments before they become embedded in infrastructural dependencies, organisational routines, market structures, and international asymmetries.

arXiv comment26 pages, 2 figures, 3 tables