The kind of game that I think is most likely to cause problems isn't the one that's simulating Formula-1 style racing or rally driving. I saw a game once (I don't know what it was called, but I expect it's a whole genre) where the players were choosing normal-people cars and racing through normal-people streets, with an emphasis on realism.
That's when I thought « If the people complaining about video games damaging people's brains were serious, they should be concerned about this much more than pretend warfare. ».
Anecdata: I'll admit to having the experience of, after playing Burnout: Revenge for many hours, continuing to view streets and traffic on my way to work through the lens of traffic-checking and drifting for extra points/boost. I hasten to add that I never felt the urge to -act- on that view; so much of driving is subconscious and/or automatic that I suspect we have something of a built-in buffer working in favor of normalcy here. But it was pretty surreal.
Maybe they make for better drivers in their prime, but as they age they would tend to have the negatives outweigh the positives. Maybe my assumptions are all wrong.
Do you know of any studies on this issue? It really does seem intriguing...
If not, why would you think video game racers are likely to make bad drivers?
Personally, I would expect faster reactions and improved machine control to both be correlated with superior driving ability (by whatever metric you choose to measure it).