The Fragile Firewall: Sovereignty, Non-intervention, and the Problem of Coercion

Abstract This article examines how artificial intelligence (AI) challenges the international law principle of non-intervention by reshaping the concept of coercion. While sovereignty guarantees a state’s freedom to determine its own affairs, the prohibition on intervention protects that freedom from external coercion. Yet, although the International Court of Justice in the Nicaragua case identified coercion as the “very essence” of prohibited intervention, it left its scope uncertain, particularly for non-forcible acts. AI-enabled operations, including synthetic media and truth-based disinformation, exploit this uncertainty by influencing state decision-making without overt force. Drawing on Marko Milanović’s models of coercion and the Tallinn Manual, this article argues for an effects-based approach that focuses on whether AI deprives states of genuine decision-making autonomy. It concludes that non-intervention must evolve to protect both territorial sovereignty and the informational domain from algorithmic coercion.