The fact that they worked at all is an indication that something similar may work in the future. It's not like apple has never had a regression in software.
Risk to the average apple customer is? Do you think that most people need to worry about this? Point Apple is making is it's not needed for the vast majority of their users and if you feel better use a piece of paper.
So let's stipulate it can happen (because well it can happen). That would take both someone being targeted en masse (prior to apple having a fix in place) and it making a difference to the person or people it happened to. Is that really a big enough risk to spend time worrying about?
