The self-driving dilemma: Safety versus freedom

A new study by the Virginia Tech Transportation Institute examines the difference between regular automobiles and the new “self-driving” models. According to the study (commissioned by an admittedly self-interested party, Google) humans behind the wheel crash 4.2 times per million miles, self-driving cars only 3.2 times per million miles.

And the self-driving era is in its infancy. As the new cars improve, pass regulatory scrutiny and gain wider adoption, tens of thousands of lives could be saved every year in the U.S. alone.

But even assuming the validity of the study’s findings, self-driving cars are not necessarily without their problems. Given government’s growing interest in controlling how and where Americans travel, they could become just another piece of our ever more pervasive surveillance state.

Above and beyond immediate, local situational awareness — staying on the road, keeping track of the distance from and speed of surrounding cars, etc. — self-driving cars need constant awareness of the larger environment: Where they are on the map, what turns to make to get where they’re taking you and whether or not there are accidents, traffic jams or road repairs ahead.

While any single piece of this information might be available from a number of sources, it’s easier to get everything from one source: A network to which the car either remains connected at all times or connects to frequently when driving. And this is a two-way street (pun intended!). The car requests information from the network and takes instructions from the network too.

This fact creates all kinds of opportunities for abuse by government agencies with command influence over the network.