Meatyfool wrote: All drivers know that stopping a car involves thinking distance and braking distance. If someone walks out in front of a human driver nearer to the car than the sum of those distances, there will be an accident. If the pedestrian is "very" near, then they may be in the "kill zone".
The same is equally true of an automated car, with the exception that the "thinking" time will be miniscule in comparison to the human driver. However, this simply reduces the overall stopping distance, and if the pedestrian steps out within that overall distance, there is still going to be an accident. The "kill zone" is still there but much smaller than before.
You're only looking at the reactive element, there is also a predictive element. Are automated vehicles as good at recognising when another road-user is (for instance) erratic, distracted or just clearly hasn't seem them, and taking some early actions like slowing down, using the lights/horn, or whatever. I bet they're not.
And here is the rub: if there is a kill zone even for an automated car, how the hell is the standby human driver going to be able to stop the car sooner, when he has to add on his own thinking distance (never mind the "this car is infallible" distance!).
By taking control before a pedestrian steps into the car's breaking distance. It may not be always be possible, but surely that is the whole point of the human?
It's true though that we don't know exactly what happened in this case, yet.