McKinsey has just published a short but interesting article on the problems of driverless cars:
It's worth a read. As they say, the technology is nearly in place, and what remains is how we phase it into our lives. They mention examples that are mostly drawn from individual human driving situations. Setting up driverless functions, like “cruse control,” in today's cars, so that a driver could turn it on or off at will. You might start using it on long highway drives or in bumper to bumper commute traffic for example.
In fact, I suspect that the first place it will be use is in business processes. Having driverless vehicles move about an industrial park, for example, or having a bus follow a standard route, for example. It's much more likely that process designers are going to insert driverless vehicles into a business process to reduce costs or increase efficiency. This article suggests some of the efficiencies that could be achieved, and it also highlights some of the risks that will need to be overcome.