We should not treat this as an acceptable strategy. If we do not have a viable mitigation for the risks of AI, then AI should be banned from public usage, just like nuclear weapons.
The fact that we don't already measure/enforce outcomes for legislative actions should tell you everything you need to know.