...The system is designed to be safe, when not damaged....
Engineers and other "Experts" offer scientific rationales for why certain disasters cannot happen.
The problem is, systems DO become damaged, they do not operate as designed, and then they are not safe.
For some unknown (to me, anyway) reason, scientists seem to eliminate the extremes of catastrophic event prediction as if the fact that that certain probabilities are so small, they can be ignored. Therein lies the peril. Those highly unlikely scenarios do in fact develop. Cf, Deepwater Horizon. Dunno, maybe it's just plain arrogance?
The current situation with the reactors in Japan is another example. Yes, there were three redundant cooling systems in the reactors, and still disaster strikes. The TEPCO representative said, problems developed that we couldn't foresee. No kidding.
So, IMO, when you hear a scientist say Oh, it's safe, nothing to worry about....yup, it may be safe, right up until it isn't...time to get the hell outta Dodge.