Academic paper identifies accountability gaps in EU AI Act for autonomous systems managing shared critical infrastructure, proposes AgentGov-SC governance architecture.

A paper submitted to arXiv in May 2026 analyzes how the EU AI Act's Annex III, point 2 excludes safety-component AI in critical infrastructure from Article 86 explanation rights and Article 27 fundamental-rights impact assessment, leaving residents harmed by combined effects of multiple autonomous systems (e.g., traffic signal and power grid adjustments on the same corridor) without a single accountable authority.

The paper identifies four residual accountability pathways—GDPR Article 22, GDPR transparency obligations, tortious liability, and NIS2—and shows each is structurally limited to individual-controller and individual-decision scope, creating a governance gap for multi-agent urban AI deployments.

The paper presents AgentGov-SC, a three-layer architecture (Agent, Orchestration, City) specifying 25 governance measures with bidirectional traceability to the EU AI Act, ISO/IEC 42001, and the NIST AI Risk Management Framework, including five conflict resolution rules and an autonomy-calibrated activation model, tested against documented UAE smart-city systems.