We are fast approaching a world where self-driving cars are both a reality and common. See for example Uber's self-driving car pilot program in Pittsburgh and Google's continual self-driving car development.
As with any computerized system, self-driving cars are susceptible to hacking (and many cases of car hacking exist already). I forsee this continuing to remain a problem long after self-driving cars are common-place.
Now, it seems to me that once self-driving cars are common-place, the need for each car to completely drive itself might become unnecessary, and instead a natural development might be that cars communicate with a master computer(s) which guides all cars on the road (while the car itself handles the simple tasks like lane-shifts and distance maintenance). The benefit of course being that you get to command a fleet of cars as a system and they can all respond to problems/accidents in tandem and traffic can be controlled by a computer with more knowledge than a single car. Whether or not such a system is likely to develop is not the subject of this question, I merely provide it as an example for my question below.
If such a system were to be hacked, it could allow the hacker to cause untold wanton destruction and death (even more so than hacking a single self-driving car). It is this fear which I believe will be a major contribution to the push-back against self-driving cars (and any system which develops surrounding it, such as the one I described above). What I'm interested in are examples of computers or systems that exist today that are feasibly hackable and would allow the hacker to cause a similar or higher level of physical destruction and/or death as in the centralized self-driving computer example I gave above. Most hacking involves stealing money or personal information, but I want an example I can point to and say: Why are you so worried about this future technology when you have a parallel technology that exists today and hasn't been exploited as you fear this new technology might? Any expansion on how hackable your example system is (e.g., what would be required to hack it and how hard it would be) would be welcome.