Driving mad

Will it kill us all?
Whenever self-driving cars come up in a conversation, there seem to be two camps: one side very enthusiastic about the new technology, and the other quite pessimistic about its ability not to crash.

As I am a performance driving instructor and an opinionated person myself, I could not pass up an opportunity to make my views on self-driving cars known to the reading public.

While most people believe that they are good or even above-average drivers, the reality is that most people don't have the reaction and instinct required to respond to an emergency with correct inputs and in time. My personal driving holy text "Inner Speed Secrets" summarizes it like this:

Of the tens of thousands of drivers we have instructed in skid control techniques, everyone would fit into one of three groups. When trying to control a skidding vehicle, they either:
  • Naturally steered in the correct direction and avoided a crash;
  • Steered in the wrong direction (most likely confused by the steer into the skid advice) and crashed; or
  • Panicked, did absolutely nothing (well, maybe screamed!), and crashed.
That means that about two-thirds of all drivers respond inappropriately in a skid situation. Believe us, it's true.

While I haven't yet instructed even a thousand drivers, I have observed the same tendencies among my students. A sizable majority of drivers will respond poorly in a bind, and they have already self-selected for having an interest in driving well!

What about those who don't even want to improve? They will still get in their cars and go to work in the morning, cup of coffee in one hand, and a bagel in another, talking on the phone, hands-free if we're lucky.

Were we to replace these drivers with computers at the wheel, I guarantee that we would see fewer accidents than we do today. Further, even the drivers who pay attention but aren't trained to respond to emergency situations correctly, will benefit from having a computerized chauffeur, as the car will deliver the programmed response consistently each time, and do it quicker than a human driver could.

The only problematic area would be situations that people writing the programs for the car will not foresee, in which case the car will turn into the equivalent of a driver from the Speed Secrets book, who does absolutely nothing and crashes. However, artificial intelligence and learning systems may help resolve that.

All in all, I anticipate the impact of self-driving cars going mainstream to be overwhelmingly positive.