This is interesting to consider. One of the reasons car centric countries make their peace with fatalities is that there is a regime and cultural expectations in place for assigning blame and imposing punishment, both criminal and civil. We know that other drivers are assholes and idiots, but there is the grim solace that if something happens they (or their insurers) will compensate us and that if it’s bad enough they may even go to jail. Furthermore, we presume that the assholes and idiots know that as well, and they will at least try not to do something stupid, especially since they could get hurt/killed too.

Given the available tech, a system relying on driverless cars is pretty much guaranteed to be safer overall, but people will resist it if there is a sense that no identifiable human is incentivized to minimize harm. If somebody gets killed in an accident and you just have Waymo (or whoever) stonewalling any efforts at compensation or justice, it becomes further dehumanizing and people will continue to prefer to take their chances with the assholes and idiots who might actually be held to account.

You come up with a regulatory regime that ensures proper insurance coverage with equal or preferably lesser friction (lord knows the American system has its issues), and also meaningful punishments to actual humans for reckless code/maintenance/routing/etc, then cultural acceptance will come.