This week, the NHTSA made a decision to designate the computer in a self-driving car the driver. This means, in theory, that you can now build a car that has no human controls in it. The only way a human could drive such a car is through some interface with the computer. This brings up a question that we have been sidestepping a lot as we develop the self-driving car: is the human driver in a self-driving car an asset or a liability? It appears to suggest the answer is “liability.” If true, the self-driving car future may parallel the move from horses to cars and, like horses now are, ban human drivers from anything but private and maybe some local roads.
Let’s talk about why this is.
The Catastrophic Failure Example
Initially, when faced with difficult scenarios, self-driving cars were supposed to relinquish control to a human driver. While this would be fine for a flat tire or a question about direction, it could be problematic if the car was traveling at speed and a near instant decision needed to be made to save the car and driver. At that point the driver, who hasn’t been driving, is given fractions of a second to comprehend the problem, determine a survivable outcome, and execute it.
In fact, even allowing the driver to be able to grab the wheel in such a case would likely result in an otherwise less damaging or even avoidable accident. The car’s computer will be able to react more quickly and appropriately.
Either the driver needs to be engaged driving, or they need to be kept out of the loop because humans don’t shift from inactive to competent active easily or quickly. And even if you could keep the driver engaged while a computer was driving their car, which is doubtful, it is still more likely that the computer could better deal with a sudden change than a human driver can. Obviously there will be exceptions but the far safer path is to design with the human out of the loop.
Do You Then Need To Own A Car?
This is the question that then comes to mind, and firms like Uber are betting the answer is a resounding “no”. Unless you live out in the wilderness (and maybe even then) it should be far cheaper and safer to have personal transportation as a service. Once you get the human out of the car you can stockpile cars for peak times, use them as delivery vehicles during non-peak times, and have them constantly roam the roads looking for riders.
Expect to see apps like Uber’s car ordering app get integrated into calendars so that your ride will arrive just when you need it and take you where you want to go. In addition expect this service to be offered as a monthly offering much like your cell phone service is so you get a certain number of miles a month at a discount and get charged for any overage.
Granted, this will make electric cars that can be charged in parking lots or while driving on roads far more attractive, because then there is no downtime for fueling up. This way, the cars are always ready to go and there is an even stronger green argument towards what will initially seem rather strange to drivers used to owning their own cars.
Wrapping Up: Benefits
The benefits of this move are going to be impressive, but they have to be, because we really don’t like change much. We move from the safety stats of cars which have around 34K deaths per year to stats that approach elevators (around 30 people a year). I could probably stop there, but we’ll also have lower costs related to car use, far more time we can spend doing fun things, gain the luxury of sleeping in (reduced) traffic, and there should be far less stress in your life unless the cars revolt (which seems unlikely).