Self-Driving Cars: It's Unacceptable for Machines to Kill Humans

With the vast majority of accidents caused by human beings, taking the weakest link out of the equation is an obvious solution. But we won't get there until we're sure about self-driving safety.

Rex Moore
Rex Moore
Apr 9, 2014 at 11:31AM

Self-driving cars will be here within the decade. Google (NASDAQ: GOOG), (NASDAQ:GOOGL) has already logged hundreds of thousands of accident-free miles with its car, and Nissan promises more than one autonomous vehicle for sale by 2020.

The self-driving revolution is expected to drastically reduce auto fatalities, and yet one of the issues holding back the adoption of this technology is... safety.

Michael Toscano, president of the non-profit Association for Unmanned Vehicle Systems International, or AUVSI, spoke with Motley Fool analyst Rex Moore about this issue. In this video, Toscano explains why driverless vehicles will be safer, but the rollout will be slow as long as these machines still have the capabilities to harm humans.

A full transcript follows the video.

Michael Toscano: Someplace between 87 and 93% of all accidents are caused by a human being. The major factor of these accidents is us, so if you take the weak link out of the system then you make the system tremendously more efficient and effective, and obviously safe.

If you can avoid having the accidents, then you can take a lot of the weight off the car. If you can take a lot of the weight off the car, then you're going to have better fuel economies and you'll have better systems.

The fact that you would not have to have stoplights, because if all the cars had the ability to communicate with each other -- this is called vehicle-to-vehicle, and also vehicle-to-infrastructure -- just like on your phone when you have a purple line that says, "I want to go from here to Starbucks," if everybody knows what their purple line is, I can make sure that I don't bump into you because I know where you're trying to get to and where you're going.

The technology is maturing and eventually, at some point in time, you will have self-driving, automated vehicles, driverless cars -- whichever term you want to use for that technology.

Moore: As Amazon will try to deliver packages with the quadricopter or flying vehicles, I'm guessing that some of the delivery companies will try to deliver with the self-driving vehicles?

Toscano: Sure -- and it may be a marsupial type of approach, where a big van comes out to a certain point and then a smaller vehicle either can fly from there or drive from there to a location.

The Amazon portrayal of what is the potential -- and that's what I think is powerful, is that this is the potential that we have now, with this technology. No one questions whether we can fly, drive, or navigate. The maturation of the technology is such that, for the most part, we can do it. Now the question is, how safe? How much can we assure that we've considered all conditions?

Now, because there is a human being, and we accept the fact that human beings make errors, that's not the same acceptance that we have for machines. It is literally in our DNA at this point in time that it is unacceptable for a machine to kill a human being -- so that's going to be one of the issues that we'll have to understand, as time moves on and as this technology matures.

That's why I say, I think in the beginning what you're going to see is those applications for which we have a higher assurance that the systems will perform as designed.