Trying to ignore your snark and overconfidence to focus on the interesting part, the new threat is a lot more people being able to use an automated tool to do something they couldn't do before.
I agree the solution is to fix the vulnerabilities. I think "impenetrable" requires a radical shift in popular mindset about operating systems, applications, and devices connected to the network.
If a model on rented hardware could manage to launch some country's nuclear weapons, the same could be done by model-armed or model-free humans, spies, etc from other countries - using their specialized tooling they could always do better than any model.
In short, AI doesn't add any new threats to the mix, the defenses against AI are the same as against adversarial humans, and the whole brouhaha is nonsense. Sure, make your defenses impenetrable, use AI for that and it'll stop any I, not only AI.