Unless, as has been suggested by many, self-driving EVs lead to a significant increase in miles traveled, which increases congestion. As much as I think they're the future and will do good for the world, it seems very likely that they will increase the amount of cars on the road.
I think people underestimate the traffic benefits of self-driving cars.
Imagine a traffic light now v. with full autonomous drivers. Now. there's at least a .5-2s delay between every car in a line starting to move, once the car in front moves. With autonomous cars, you'd probably see the entire column of cars begin moving within 2 seconds. A traffic light that's normally a two minute affair is cleared up within 15 seconds.
Rivers don't have traffic because all the water particles are moving simultaneously. With half-decent infrastructure, traffic could become a thing of the past with fully autonomous cars.
Or better yet (as some people have modelled), fully autonomous cars wouldn't even need traffic signals and could utilize much more of every intersection at any given time.
I'm curious how easily people think large numbers of people will get into a vehicle with no steering wheel, or a steering wheel with nobody sitting at it, to carry them to where they want to go? That looks to me like a gargantuan release of control on the part of customers. Granted, there are plenty here ready to ditch the (steering wheel), and there are always the innovators / early adopters. But Early Majority and crossing the chasm into widespread acceptance?
Cost of ownership and insurance are likely to be two very huge factors pushing people into acceptance. Cars are pretty expensive to own--second only to houses, but unlike houses they whittle away and depreciate into nothingness pretty quickly. A house on the other hand, when you have a 2-4 people equally sharing living costs, ends up being a relatively minor expense and appreciates with time.
With the car2go carsharing service in my own city, I calculated that I could drive about 4000 miles annually on the carsharing service before having my own vehicle is worth it. Now imagine such a carsharing service when the fuel is a fraction the cost, where the cars go from maybe 10-20% utilization to 70-90%+, all the cars are low-maintenance and long-lasting electrified drivetrains, and when there's no crews needed to fuel up and go out to maintain vehicles sitting at the side of the road. That 4000 mile number could very easily and very quickly jump to 20,000-50,000 annual miles needed before a personal car is more worthwhile, as the level of autonomy for the cars go up.
Look at the popularity of services like Lyft and Uber, which could be slashed in price with fully autonomous EVs.