National AI kill switch is ineffective

The UK government's refusal to back a legal 'kill switch' argues that a single country cannot reliably switch off dangerous models because the same systems can be developed, hosted, or accessed outside its jurisdiction. This shifts attention from unilateral shutdown powers to international coordination, export controls, platform choke points, and resilience measures for national defense and critical infrastructure. — Frames a practical pivot in AI‑safety policy: nations must prioritize cross‑border governance, supply‑chain interventions, and resilience over standalone 'turn‑it‑off' laws.