Structural Preconditions of Algorithmic Sovereignty: Informational Fragility and the Institutional Architecture of AI Governance

Review

Hervé and Mbenga recast AI governance as a state-capacity problem. Their central concept, informational fragility, describes the inability of public institutions to produce reliable, interoperable demographic intelligence. The GPAS framework then treats algorithmic sovereignty as an emergent property of three aligned capacities: Structural Integrity, Institutionalized Algorithmic Capacity, and Governed Interoperability. This matters because it moves the governance question upstream from model regulation to the institutional substrate on which public-sector automation depends.

The paper is especially useful in distinguishing technical interoperability from governed interoperability. It recognizes that integrating registries redistributes informational authority, can trigger bureaucratic resistance, and can centralize power even while improving administrative coherence. The DRC case and prototype make this institutional geometry visible, while the proposed sequencing rule, stabilize information, establish governance, then expand analytics, is a practical corrective to procurement-led digital modernization.

The methodological limit appears in the Algorithmic Sovereignty Index. Birth-registration completeness, the UN Online Service Index, and a three-level data-protection score are tractable proxies, but they do not directly measure the institutional properties the framework claims to operationalize. The index also compresses legitimacy into formal legal and oversight indicators. The authors explicitly state that ISA does not measure democratic quality or human-rights compliance, which is a defensible boundary for an exploratory index. Yet the paper also describes algorithmic sovereignty as a legitimate property of public authority. That stronger claim requires evidence about contestability, redress, enforcement independence, abuse constraints, and whether affected people can correct state-held demographic data.

The resulting governance tension is productive rather than fatal. Informational coherence can increase administrative capacity while also making surveillance, exclusion, and executive coordination easier to scale. My assessment is that GPAS currently measures the administrative preconditions for algorithmic sovereignty more convincingly than it measures legitimate sovereignty itself. The distinction matters because a state can become informationally coherent without becoming more contestable or institutionally constrained. A next iteration should test whether the sovereignty claim survives cases where coherence is high but correction, redress, and institutional independence are weak.

Key Insight

GPAS measures the administrative preconditions for algorithmic sovereignty more convincingly than it measures legitimate sovereignty itself, because informational coherence and state capacity do not establish contestability, institutional independence, or justified public authority.