Algorithms of Evasion: The Rise of AI-Enabled Proliferation Financing

RUSI Research Papers
10 juin 2026, 02:00

Texte de la source originale

This paper explores how AI is revolutionising proliferation financing and sanctions evasion, revealing new threats, regulatory gaps and actionable policy solutions.OverviewThis paper explores how AI is revolutionising proliferation financing (PF) and sanctions evasion, particularly by states such as North Korea and Iran. The report warns that a shift from AI-assisted to autonomous, 'AI-enabled' evasion tactics, including the mass production of fraudulent documents and optimised cryptocurrency laundering, threatens to overwhelm existing manual and rules-based detection systems.Key RecommendationsNational authorities should explore 'safe harbour' provisions for AI-driven counter-proliferation models, establish 'compute-KYC' and cloud liability standards, and create a National Security Data Enclave to pool pseudonymised transaction data.The Financial Action Task Force should update Recommendations 1 and 15 to address vulnerabilities from autonomous AI agents, while global bodies should develop 'know-your-API' standards.Financial institutions must upgrade KYC procedures to defend against deepfakes and deploy defensive AI to audit trade documentation for semantic inconsistencies.