Normative Inflation and the Crying Wolf Effect in the International Governance of AI

The international governance of artificial intelligence (AI) is often said to suffer from a regulatory gap. In this article, I argue that the opposite problem is equally significant. The governance field is characterised by institutional proliferation and normative inflation: an excess of principles, declarations, and advisory bodies that lack binding authority or hierarchical coordination. This 'crying wolf' dynamic results from repeated, high-urgency normative signalling that, without enforcement mechanisms, gradually diminishes the credibility, relevance, and guiding power of the norms themselves. I distinguish between iterative soft law-which advances towards binding instruments through successive approximation-and substitutive soft law, which replaces enforcement with expression. I contend that this dynamic poses a structural threat to the performative credibility of technology governance and represents an ethical failure, as it gives everyone, particularly those most vulnerable to AI's harms, the false illusion of protection while offering none. The challenge facing the international governance of AI is to cease generating more empty signals and to strengthen institutional influence through delegated authority, normative and epistemic integration, and credible consequences. I conclude by proposing three design criteria that any future consolidation effort must meet, and by asserting that the ethics of governance architecture and its design deserve as much philosophical scrutiny as the ethics of the technologies being developed and overseen.