Why waiting for autonomous car perfection could be a deadly mistake

7 Nov 2017

Car crash. Image: PongMoji/Shutterstock

According to a new report, even if autonomous cars were just a little bit better than humans, it could save countless lives.

One of the things that stymies the launch of commercial autonomous cars is the fact that manufacturers and designers need to show that they can get from one point to another without putting humans in harm’s way.

However, a new report from the RAND Corporation – a US government think tank – has claimed that even the use of moderately capable self-driving cars would go a long way towards reducing the number of road deaths.

The report estimated that even if autonomous cars were just 10pc better than US drivers, it could prevent thousands of road deaths.

Estimates put the number of US road deaths at more than 40,000 last year. If we were to wait until autonomous cars were between 75pc and 90pc better than humans, hundreds of thousands of lives might be lost unnecessarily.

By the time 2070 rolls around, an estimated 1.1m lives would be saved once autonomous systems are up and running.

Perfection is the ‘enemy of good’

“Our work suggests that it is sensible to allow autonomous vehicles on America’s roads when they are judged to be just moderately safer than having a person behind the wheel,” said Nidhi Kalra, co-author of the study and director of RAND’s San Francisco office.

“If we wait until these vehicles are nearly perfect, our research suggests the cost will be many thousands of needless vehicle crash deaths caused by human mistakes. It’s the very definition of perfect being the enemy of good.”

To come to this conclusion, the RAND researchers’ calculations were made by estimating road fatalities over time under hundreds of different plausible futures and different safety requirements for autonomous vehicle introduction.

A PR battle

The greatest challenge to this idea is essentially a public relations one, RAND said, with the reality being that autonomous cars will still be involved in crashes, but people’s perceptions of them will be more critical than a crash caused by a human driver.

“This may not be acceptable because society may be less tolerant of mistakes made by machines than of mistakes made by people,” said David Groves, another of the study’s co-authors.

“But, if we can accept that early self-driving cars will make some mistakes – but fewer than human drivers – developers can use early deployment to more rapidly improve self-driving technology, even as their vehicles save lives.”

RAND hopes that the findings will be enough to persuade some members of the US government to open up regulations to autonomous car developers. Currently, the likes of Google and Intel are only now getting to test these cars on public roads.

Car crash. Image: PongMoji/Shutterstock

Colm Gorey was a senior journalist with Silicon Republic