I'd argue that in virtually all other forms of engineering, our risk tolerance are reflections of our technological capabilities at the time. We now have the technologies to make buildings, electronics, planes and all kinds of other wonders that don't kill their users, and for that reason,
we have a social expectation that these things should not kill their users.
It wasn't always like this. In the 50s, people would've been a lot more tolerant of airplane deaths, for example, because the technology at the time couldn't reasonably accommodate better safety standards. I could say the same about building standards in the 1850s. But as technology got better, we expected better, and many millions of lives have been saved as a result.
For some reason, that never happened with cars. We've long had the technology to build cars and roadways that don't kill people, yet our social expectations have not reflected that technological achievement.
So yes, it's true that risk levels we choose are never zero. But in the case of our roadways, the risk levels we've chosen is far higher than it should be, given our level of technological achievement. We can get it closer to zero.
I know we all support these principles (why else would we be here on a Friday night
), but I think this attitude that "everything has risks" is a dangerous one. While it's true, this attitude can too easily be used to argue that everything has risks, so there's no point in doing better. This line of thinking has been used to argue against virtually every proposed safety measure ever.
I much prefer the attitude that zero road deaths are acceptable, and that every road death must be treated as a critical engineering/policy failure (which is precisely what it is).