City-Level Sovereign Artificial Intelligence as a Valid Global Digital Compact Target: Topology, the Digital Agora and a Verifiable Cryptographic Stack for Localising the Compact
The Global Digital Compact, adopted by Member States in September 2024 as an annex to the Pact for the Future, sets shared objectives on digital divides, the digital economy, human rights online, data governance and the international governance of artificial intelligence. Whether those commitments matter depends on what happens in cities, where public services, infrastructure and community life meet digital systems. This paper argues that a city-level deployment of sovereign artificial intelligence is a valid and auditable target for the Compact's localisation, provided that democratic agency is written into the protocols that govern the system rather than left outside them. It brings together four bodies of work developed across a single line of inquiry. First, the topology of city-level AI: governance under Permanent Systemic Instability shifts from territorial monopoly to Flow Sovereignty, the maintenance of coherent, low-surprise trajectories across digital, physical and financial systems, organised through active inference, information-geometric policy navigation, sheaf-theoretic consensus and a 72-hour autonomy standard. Second, the Digital Agora, the rights-based counter-model to the Algorithmic Leviathan, which relocates democratic constraints from the political layer to the protocol layer. Third, the primitives that make such constraints enforceable: layered Sybil resistance, quadratic weighting of preference intensity, and the Minimal Anti-Collusion Infrastructure (MACI), which makes votes private, receipt-free and publicly tallied. Fourth, the cryptographic foundations of verification: zk-SNARKs, the per-circuit trusted setup of Groth16, the universal and updatable setup of PLONK, and Kate–Zaverucha–Goldberg polynomial commitments. The paper composes these into a six-layer deployment architecture, maps it to the Compact's five objectives, states five validity criteria a city deployment must meet, and sets out a pilot programme whose outputs are the local evidence the Compact asks for. It also states the limits: residual trust in setups and coordinators, the trade-off between Sybil resistance and inclusion, prover cost at city scale, and the sovereignty trap of new external dependencies. The work is a design specification; its single computation is an analytic illustration and contains no empirical data. KSCCN article prepared in connection with the GDC Localization Townhall, "Localizing the Global Digital Compact: From global frameworks to local implementation" (UN Office for Digital and Emerging Technologies, 18 November 2026). Companion to "City-Level AI: The Topology of Command and Information Geometry (Volume II)" (DOI 10.5281/zenodo.22990597) and "The Algorithmic Leviathan" (Volume I, DOI 10.5281/zenodo.22990607). Files: full-colour KSCCN PDF, editable DOCX, and a 14-slide presentation deck (PPTX).
Authors
- Syed Muntasir Mamun (ORCID: https://orcid.org/0000-0001-6845-2853)
Institutions
- Ministry of Foreign Affairs, Dhaka (BD)
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-11-18
- DOI
- https://doi.org/10.5281/zenodo.24342653
- Primary Topic
- Smart Cities and Technologies
- Type
- article
- Field-Weighted Citation Impact
- 0.00