Individually, you and I can both agree that the bar is quite low for self driving cars. It just needs to do better than current drivers to be safer. We should adopt it immediately, if we can just prove that it crashes less.
Society doesn't see it that way. 1 crash is horrific. 1 accident, is proof the technology is flawed. 1 accident is 1 too many.
It's stupid, but that's how it is.
The question of liability is always tough to resolve. Is Boeing responsible for the 737 MAX deaths because they wrote the faulty software? Is the airline responsible for poor training or buying the version without the "disagree" light? Or is the FAA liable for their poor oversight of the whole program? Arguments can be made for all three.
Lane-assist on the highway? Sure.. autopilot? Not any time soon. Probably not within 20 years
I don't have any sources to back this up, but I'd wager that people, on average, would prefer to take a risk an order of magnitude greater if they believed they had some degree of control over the outcome.
I.e. let's say 10% of conventional cars experienced a fatal accident, and 1% of self-driving cars experienced a fatal accident. I think most people would choose to be in the 10% fatality group because they feel like they would be in control.
Anecdotally, I have a co-worker who hates flying because she feels like she isn't in control, even though aviation deaths are extremely rare.