I agree. Less margin for error, but leaves people who depend on automation vulnerable. I just imagine lots of growing pains before we get to ideal state.

I don't know about your city, but I trust technology a lot more than the average driver.

I don't. Technology can be subject to glitches, bugs, hacking, deciding to plow right through pedestrians (hello Tesla!), etc.

While the case can be made that human drivers are worse at reaction time and paying attention, at least a "dumb" car can't be hacked, won't be driven off the road due to a bug, won't try to knock people over itself without stopping, etc.

A human, when they catch these things happening, can correct them (even if it is caused by them). But if a computer develops a fatal fault like that, or is hijacked, it cannot.

EDIT: It seems like this community is full of AI techbro yes-men. Any criticism or critical analysis of their ideas seems to be met with downvotes, but I've yet to get a reply justifying how what I said is wrong.