It's been like a year or so, and things might have changed, but I remember reading in a blurb somewhere that both Toyota and GM were looking at not releasing self driving cars until they could do so without having any human backup in place. For the arguments you make, one, by the time an emergency happens the human isn't ready to take over, and b, there's a legal grey area in that.veen could give you chapter and verse on this, but there are different levels of autonomy and everybody but google is going for the fuzzy middle where people are supposed to intervene if things go pear-shaped. Google argues (correctly, I think), that this is exactly the wrong way to go about it - because now you're not trying to determine whether you're safe or not, you're trying to determine if you know better than the car or not.