Unified Digital Government Platform with an AI Core: Reference Architecture, Risk-Classified Human Oversight and an Open Economic Model (Technical note v1.0)

Technical note recording a country-agnostic reference architecture for a single government platform with an AI core: a shared data space based on the once-only principle, interoperability between agencies, and an AI core that prepares explainable recommendations while decisions and accountability remain with officials. The note defines four risk classes that set the level of algorithmic autonomy, human controls and audit for each AI-supported decision, and documents an open, reproducible Monte Carlo model of the platform’s macroeconomic value. Code is included (Apache-2.0).