From Poles to a Continuum: The Explicit Delegation Module as a Response to Critiques of Human-in-the-Loop Architectures

This paper presents an evolution of the decision-making protocol architecture introduced in “Synthesis: An Architecture for Goal-Oriented AI in Crisis Governance” (DOI: 10.5281/zenodo.18098419). In direct response to substantive public critique—which exposed a fundamental polarisation in how automation is perceived, ranging from outright rejection of algorithmic decision-making to expectations of full AI autonomy—we propose a new structural component: the Continuum Delegation Module (CDM). The CDM is designed to formally institutionalise a spectrum of delegation levels, explicitly defining how authority is shared between humans and algorithms within a single governance system. It enables the design and operation of adaptive governance systems in which, for each process or phase, it is unambiguously specified which operations are executed automatically by deterministic rules and which remain under human control. The module thus addresses not only technical objections regarding the limits of AI capability but also provides a transparent mechanism for building public and expert consensus on a core societal question: “What, to what extent, and under what safeguards are we willing to delegate to algorithms?” This article outlines the concept, its architectural integration—including synergy with the Ethical Audit Module—and illustrates its application through a hypothetical case study on reorganising emergency medical dispatch services.