Is the cost to society (in lives lost) worth the cost to individual enjoyment?
Is a collective devaluing of human life worth saving a few individual lives?
Consider if we cut down all trees within a dozen or so meters of roads, and installed guardrails everywhere. We could probably save many lives, but at the cost of devaluing the beauty of the countryside. Everyone's lives would become just that much more depressingly monotonous and sterile, for the sake of a preventing a few roadside deaths.
Safety should not be our society's primary concern. Quality of life is a much better thing to strive for.
You realize that we did that and it resulted in more deaths, right?
We could string safety nets between all skyscrapers and under all bridges. We could ban convertibles. We could put fences along all rivers and ban backyard swimming pools. We could ban fireplaces.
There are any number of things that we could hypothetically do to prevent accidental or untimely deaths that we do not do because they would degrade the quality of our lives.
Skyscrapers generally don't have windows that can be opened or easily broken, and very rarely have balconies. When's the last time you heard of someone going through a skyscraper window? It's a safety net.
We probably couldn't string safety nets under all bridges. However, we have already been fencing them off to prevent jumpers.
I'm not clear on why you are targeting convertibles specifically; I'm unaware of any special danger they pose. However, I would argue that banning cars entirely would raise our quality of life, if combined with other measures.
I'll concede the river one since there are too many possibilities there to argue exhaustively.
I could prevaricate on the swimming pool thing, since my parents own one and our quality of life probably would be better off without it, but not worth it.
I'm not clear on enough building codes to know how much we discourage fireplaces, but most new buildings don't seem to be built with them. We appear to have invented central heating over the last several decades, which is both safer and more effective.
The point being that the reason we don't implement a lot of "safety" measures is not because they "degrade the quality of our lives" but because they don't work. In fact, most of the so-called safety measures you luridly imagine actually should be done because they'd increase the quality of our lives... so really, the question is why we haven't done them, if QoL is such an important factor.
It depends on how much pleasure you get out of driving. I certainly think it's worth some fraction of a life. And of course, a certain amount of the risk comes from inattentive drivers who don't enjoy driving, and would rather read a book or eat cereal (or both) on the road.
What I'd like to know, though, is that as self-driving cars become developed, the labs are calculating the risk factors of various intersections, streets, exchanges, etc. That is, different roads can be analyzed by the robot cars and independently assigned a risk factor for a given section of street.
The resulting impact to traffic engineering and city resource prioritization could be a very interesting side-effect to the introduction of self-driving cars.