It’s not about everything being automated. We also have to differentiate between early incarnations of autonomous vehicles and the desired, final state.
A manual override will of course be part of early models for the foreseeable future, but the overall goal is for the machine to make better decisions than a human could.
I don’t have any quarrel with life or death decisions being made by a machine if they have been made according to the morals and ethics of the people who designed the machine, and with the objective data that was available to the system at the time, which is often better than what would be available to a human in the same situation.
It’s the interpretation of said data that is still not fully there yet, and we humans will have to come to terms with the fact that a machine might indeed kill another human being in a situation where acting any different might cause more harm.
I don’t subscribe to the notion that a machine’s decision is always worth less than the decision of another entity participating in the same situation, just because it so happens that the second entity happens to be a human being.