I think there are reasons to be a bit more optimistic about the dangers.
A) The reason that cars crash (crash meaning doing damage in the process of failure) is by a large margin not mechanical failure, but human error. While in this case the technology is still infantile and will, even after a couple of years of testing, certainly be more prone to mechanical failure, I see no real reason to assume that it would ever be permitted for actual inner city usage before it beats a certain range/failure-ratio, because that seems trivial to secure and exactly what regulatory bodies are made for.
B) Let's assume there will be no humans piloting these things in an urban transportation scenario, ever, because that does indeed seem incredibly dangerous and silly. I don't think we as a species have the capacity to manually navigate these things in a 3d space filled with them. I don't see any reason we have to. Let's not do that.
C) If you think about it, physics are impossibly bad for car safety. There is no time to decelerate safely and then also no safe space to evade to. In contrast, in the case of flying thingies however, you have space and time working for you: Even if parts of the machine fail, it definitely has at least an order of magnitude more seconds to manoeuvre to relative safety. Of course, catastrophic failure (big fireball in the sky) can possibly still occur (I assume), but you are not working against physics, which is great, because then all you need is outstanding engineering and appropriate fail safes. Engineers will take that over having to beat physics any day.