Much of the promise of autonomous vehicles hinges on their potential to cut road deaths. And even before the technology is perfected, they could save thousands of lives each year, according to a new report, which suggests that they should only have to be moderately better than human drivers before they are widely used in the United States.
The report by the RAND Corporation – The Enemy of Good: Estimating the Cost of Waiting for Nearly Perfect Automated Vehicles – suggests that allowing the widespread use of autonomous vehicles when they are just 10% better than current American drivers could prevent thousands of road deaths over the next 15 years.
Within 30 years, found researchers, that number could even reach hundreds of thousands, compared to waiting for the vehicles to become 75% or 90% better than human drivers.
“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.”
As autonomous vehicle developers test their cars, lawmakers are assessing the regulatory changes required to support their deployment. But it is not yet known how safe the vehicles will have to be before they are made available for wider consumer use, note the authors.
Even if autonomous vehicles are proven to be safer than the average human driver, the vehicles would still cause crashes, believe researchers, due to hazards such as inclement weather, complex traffic situations and even cyber-attacks.
“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, study co-author and co-director of RAND’s Water and Climate Resilience Center.
“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.”