As with anything, there are a few roadblocks in the way.  Driverless cars need to be able to demonstrate they can handle the worst of human driving, which seems feasible but has yet to be fully tested. Humans are much better than robots at handling sudden changes and improvising, although roboticists are hard at work improving that.

The majority of cars would need to be driverless, and that’s a much trickier proposition. It takes time for any new technology to filter down to the general public, especially in America, where the average age of a car is over a decade and just keeps climbing. And currently, driverless cars are an expensive proposition; Google’s driverless Priuses cost $150,000, and that is just for the equipment to make the car driver-free.

Then there’s the issue of trust:  How many drivers will believe a computer can drive to the store for them? And how many would prefer to have control over their vehicle? Cost doesn’t really matter if nobody will buy the car in the first place.

It seems more likely that technologies that come out of driverless cars will come onto the market over time, and that driverless vehicles will become more likely on a large scale, such as in public transit. Vehicle-to-vehicle systems, which coordinate drivers and warn them about oncoming accidents, hazards, and drivers with problems on the road, are currently being tested by the federal government. And limited driverless technology is in the works, handling tasks such as parallel parking.