I suspect that there would also be some negative effect of consistently increasing the cognitive load associated with driving, as opposed to looking at performance on specific roads on which a given driver probably only spends a small portion of their total drive time. At some point, I would expect the detriment of mental fatigue to outstrip the benefit of "forcing" people to pay attention.
When I drive and see a huge truck driving at high speed from the opposite direction, I tell myself "just keep your lane, he will keep his lane too, and it will be fine" (and I probably unconsciously slow down a bit). Without the lanes I'd probably feel a bit less comfortable.
On the other hand: having visited a number of EU countries, IMO the UK seems to have 2-3x more clutter on the roads (lanes, texts for drivers and pedestrians and what not) than any other of them. They have texts everywhere!
This is an interesting idea. Maybe we should be looking at this as an optimization problem where drivers have a limited budget for attention and we want to concentrate it in the parts where it's most useful by varying the quality of UI assists?
