The whole video call when driving is intriguing from a Human Factors POV. Drews and Strayer have done alot of research on the effects of cell phone communication (wireless and holding handset) and also just having a passenger near you. The performance for driving while talking on a cell phone (or even texting) has been comparable to the performance of a drunk driver. Some people have argued that the difference between having someone in car versus someone on the phone is that the other individual is assisting with the driving or shuts up when you need to dedicate extra attention to a driving task.

Theoretically if that's the case then it could be tested in a driving simulator, with the front facing camera focusing on the driver's face, or the rear of the camera (if using an iPhone4) on the actual drive could result in less driver error (swaying through lanes, not following posted speed limits, etc..)

Though the university I'm at doesn't have a realistic car driving simulator (though GMU does! It's cool!). Heh, maybe if I decide to get my PhD and I apply to GMU..