Driverless Cars: Beyond the Hype

Driverless cars promise get people around more safely. However, the basis for that assumption is often backed up by old statistics that were not considering the safety of autonomous vehicles.

In fact, not long after I submitted my Actuarial Review article about driverless cars, “60 Minutes” presented a segment, “Hands off the Wheel” on the same subject.

Since I had intensely researched the topic, I could not wait to hear what the reporter would tell the general public. Instead of investigative journalism, the segment gave the driverless car industry a boost with little mention of the many unresolved issues and potential unintended consequences.

At its beginning, the reporter said “almost all” car accidents are caused by driver error, noting the safety potential of driverless cars. The truth is, nobody really knows how safe driverless cars will be.

The often-quoted statistic by driverless car advocates is that 93 percent of car accidents are caused by human error. The logic is that by reducing the opportunities for driver mistakes, automated vehicles will be safer.

The statistic and its assumptions, which were also presented as testimony before the U.S. Congress, are very important because they guide the assumptions and expectations of driverless cars. Google also boasts that all of the accidents involving its cars were due to human error.

But when the rubber hits the road, it’s the insurance industry’s opinion that counts. Its actuaries not only have the most experience looking at the factors that lead to accidents, but the industry will be responsible for covering them.

My article, Destination Driverless: Will Vehicles – Not Drivers – Become the Center of Risk?, sets the record straight about the all-important 93 percent statistic thanks to actuarial analysis provided by the Casualty Actuarial Society’s Automated Vehicles Task Force.

The task force concluded that automated technology could only address 78, not 93 percent of accidents if it cannot overcome factors such as weather, vehicle errors and inoperable traffic control devices. Using the 93 percent statistic, the task force also asserts, is problematic for other reasons.

Stemming from a National Highway Transportation Safety Administration (NHTSA) study, the statistic had nothing to do with driverless cars. And, due to its age, the study could not anticipate the latest safety improvements to conventional vehicles.

So what do actuaries need to have a better idea of the potential costs of insuring driverless cars? Access to proprietary data that developers and manufacturers naturally are not quick to share.

My article also details other factors that should be considered – especially when human drivers must take the wheel of automatic vehicles. It also covers the challenges that developers must overcome to make them viable in the real world.

What does this mean to the average consumer? Excitement about driverless cars abounds, but nobody sees a significant population of driverless cars for another 20 years.

In the meantime, we can expect driverless cars to gradually join the traffic jam. That transition, in and of itself, could also lead to unintended consequences.