When?  I don't know but it is remarkable how far the scientists and researchers and engineers have come in just a few short years.

DARPA held the first Grand Challenge in 2004.  An off-road course to challenge a fully autonomous robotic car.  Not one car finished.  Most failed within the first few miles.

DARPA ran the same challenge a year later.  5 cars finished and most of the rest of the field improved on previous results.

This year DARPA ran a new challenge, the Urban challenge.  Designed to replicate an Urban experience.  It was won by the team from Carnegie Mellon University.

Tether said Tartan's vehicle averaged about 14 miles per hour throughout the course, which covered about 55 miles. Stanford averaged about 13 miles per hour, and Virginia Tech averaged a bit less than that. In response to a question from the press, Tether said that MIT came in fourth place.

Remarkable. Carnegie Takes First in DARPA's Urban Challenge.

I had read about efforts to automate car driving years ago.  Typically it involved very expensive retrofits to our existing roads to provide the guidance to the cars.  That was before the computational power we have now was available and before GPS was so widespread.

The computing power now takes up the entire back of a Passat station wagon.  The sensors are ungainly attachments bolted on the roof.  This will change, this will shrink.  It will become ubiquitous and invisible and accepted and expected.

You will be able to get in your car, type in your location to Google Maps.  turn your seat around and surf the web or perhaps chat with a co-passenger.  Because there won't be a driver anymore.  Why would you want to drive a car?  That's so 2007?