But, they are quite there yet. You're not asking a car to do a cross-country race, you're asking an AI the widely studied and even perfected question "¿How do you get from point A to point B in the least amount of time without hitting anything?". Sure, reading about a pedestrian being killed by an AI is unsettling. But is simply not the machine's fault. Perhaps the note should read "Pedestrian killed by error in self-driving AI programming". The humans behind the screen are at fault, not the car itself. Tesla is doing wonderfully with their self-driving cars because they're AI team is wonderful, and you'd be surprised at just how much AI's have consitently beat humans at these sort of tasks, both in speed and safety.