Journal of Interdisciplinary History and Human Societies
From Meaningful Human Control to Decision-Time Accessibil-ity The Architectural Conditions of Responsible Oversight in Automated Systems
Abstract
Jean Claude Havyarimana
Meaningful Human Control (MHC) requires that the operation of an automated system be traceable to a human agent in a position to exercise the capacities required for responsibility. Tracing demands more than formal authority; it demands that those capacities remain accessible at the moment of action. This article names that requirement decision- time accessibility (DTA): the architectural condition under which responsibility-relevant capacities reach occurrence under institutional and technological load. DTA is derived directly from MHC’s tracing requirement and decomposes into six elements: Truth, Relation, Integration, Doing, Orientation, and Attention & Agency. Where the time required to assemble these elements exceeds the decision window the architecture provides, the system has crossed what the article calls the Judgment Cliff. The article develops the framework through three retrospective cases (Challenger, the Boeing 737 MAX, the 2008 rating-model chain), set against Crew Resource Management as a counter-case, and applied prospectively, with prospectively stated predictions and explicit disconfirmation criteria, to algorithmic clinical decision support under the EU AI Act’s Article 14. Where DTA collapses, operator-blame allocation becomes fictive accountability: a misattribution of architectural pathology to individual fault. Meaningful control requires architectural design, not authorizing structure alone.
