Driverless vehicles probably do have the potential to decrease accidents and allow better use of travel time, but I wouldn't bet on them reducing commute times. To the extent that they work to lower commute times with a fixed population of commuters, that will draw more commuters until a new equilibrium is reached. Studies of road building usually find that the equilibrium commute time reached is equal to or higher than the old equilibrium.

We may have a fighting chance if we ever get economic short range vtol air travel. Instead of a 20-50% increase in capacity like adding lanes (and probably driverless vehicles would be in this range as well) or a 5-10x increase (like building a highway), that might get us a 1000x increase in capacity. That might be big enough to move the equilibrium somewhere interesting.