It doesn't really matter whether autonomous cars crash less than human drivers. A computer can never be punished, so a computer must never be allowed to kill any humans.

How does punishing someone make crashes better? I would rather have fewer crashes because of a system that can't be punished than more crashes with punishment.

Being able to punish someone doesn't change the fact that a crash happened and it doesn't change the fact that someone died. There is a difference between punishment and justice, there is rarely ever perfect justice in life.

The best we can do is mitigate harms as much as possible, and it seems self-driving might be the best mitigation of traffic accidents compared to human drivers.

Suppose human drivers were killing 1 person every 100 miles driven, and software was killing 1 person every 10 billion miles driven. We should just accept that the human population will be decimated each day for accountability reasons?

I do think that currently we're at the stage where Waymo is trying to convince most of the world of the _quality_ of its offering, so for the moment, things are well aligned.

But in 5 years if there are multiple self-driving offerings competing on things like price, door-to-door time, OPEX (e.g. does a camera-only system with no moving parts need less skilled maintenance?), then someone may try to be the cheap but good enough self-driving service that cuts corners -- and we should want some execs to actually go to prison rather than just some cost-of-doing-business fines.

Where do you live that humans are punished for "accidents"? You can run someone over in America and it's basically an "oopsie, do better next time!"

* 12 deaths per 100,000 per year, but 4 people go to jail for it

* 2 deaths per 100,000 per year, but no one goes to jail

Because apparently going to jail is important for some reason?

Honestly, choosing jail is somewhat dystopian IMO