My main argument here is decision making. A computer thinks a car is a car. It thinks a person is a person. But a computer doesn't have the rational decision making humans have. There are some situations in which a car has to crash. The computer will not be able to tell the difference between a 10 million dollar bugatti and a 10,000 dollar volkswagon. A computer wont be able to tell the difference between an old lady, a baby, and a young man. A computer wouldn't make the rational choice to hit the Volkswagon.
Not only have you assumed that a situation where a crash is imminent between multiple self driving cars (if self driving cars really are the future), but you have also stated that the computer will not be able to tell the difference between an expensive car, a cheap one, and a splattering of different pedestrians. I argue that first of all, the fact that a 10,000 dollar vw with self driving capabilities is more than unlikely. Secondly, self driving cars currently have a type of network where they can interchange information between each other. If we already have a system that can avoid your worst case scenario, what will a future full of self driving cars yield?
Sure the technology can be used to avoid these situations, but all technology malfunctions without a fail. You ask the question what will a future full of self driving cars yield? Well for one, MIllions of hackers. We are essentially handing MILLIONS of people to criminals on a silver platter. All they have to due is hack into the car, which has been done previously, and they are able to kill millions of people at a click of a button. Going back to the point of technology malfunctioning, what guarantee do we have technology won't fail us, or Hackers wont get to us first?
Yes, although the risk of hackers is a major problem of self-driving cars, the hacking of a self-driving car was only to show the possibility of a hack. The car manufacturers certainly weren't paying close attention to this possibility because of the recent arrival of self-driving cars, and put little cyber security onto the code. Now that the possibility of car hackers has opened up, self-driving car developers will certainly make cars more secure.
You mention that millions of hackers will be prominent in the future, and while there will be many, only a handful will be successful. You also say that hacking ONE car will grant access to kill millions at a click of a button. The process of killing millions of people at the click of a button from only one car is very unrealistic. By the time a car is hacked (which will be extended due to developers increasing security), a warning would be issued to other self-driving car owners. Hacking a car and a network are entirely different things.
Yes while it may be difficult for hackers to get into a vehicle it doesn't mean they won't. You still haven't mentioned anything about technology malfunctions that are unavoidable. Sure, normal cars malfunction, but a lot less considering that they have been invented since 1886. Also you are suggesting hacking into a single car is a small ordeal. But the amount of deaths due to a hacking of a single car will certainly add up.
Cars are very unlikely to become the future because of how hard they are to implement into society. Many people in the poor areas of the world don't have normal cars. To get most of the population using self driving cars is an unrealistic goal.
On a side note, This was an awesome debate. Thanks for participating.