The recent death of a man who perished while behind the wheel of a self-driving car highlights that when it comes to the open road, human error remains the biggest risk, say insurance experts.
Joshua Brown of Canton Ohio, 40, died when the censors on his Tesla Model S failed to differentiate the bright white sky from the side of a turning truck. His vehicle, which was set on autopilot, failed to stop for the oncoming truck, shearing off the car’s roof, and continued for some distance down the highway.
However, recent reports allege both human drivers failed to do due diligence while on the road; a DVD player was found within the Tesla, with several witnesses stating a movie was still playing at the scene of Brown’s death. The driver of the truck also had a long history of safety citations, with seven reported during four traffic stops over a two-year period. Both oversights underline the gap in current technology and human use, says Barrie Kirk, co-founder and executive director of the Canadian Automated Vehicles Centre for Excellence.
“Human beings make a terrible plan B,” he says. “The driver should have been alert. Clearly all of the companies doing testing emphasize the human must be able to take over if necessary, but that’s really difficult to do.”
But who should be responsible to properly preparing drivers for a hands-off reality? Paul Kovacs, executive director of the Institute for Catastrophic Loss Reduction and author of the Insurance Institute of Canada’s
Automated Vehicles Research Report, says the onus should be on insurers to both educate drivers, and ensure automated products are priced in a way that reflects the true risk.
“The insurer must understand what it is that the driver knows and doesn’t know,” he says. “How far along is their learning and awareness? That’s a critical part of providing insurance, and it’s not straight-forward.”
Both Kovacs and Kirk agree that the accident, while tragic, won’t derail the progress of bringing autonomous vehicles to market, and learning how to ensure them.
“The message is that the technology is still moving forward, but not there yet,” says Kovacs. “Tesla is not offering a totally autonomous car today – even if the driver thought it was safe to get it go off on its own. It is not yet there, and insurance companies do have to watch, because these cars not having accidents is simply not true.”
Kirk adds that in the case of the vehicle’s sensors, the technology is designed to prevent only rear-ended accidents, and won’t be suitable to identify forward or sideways-moving threats until 2018. “The technology is being developed in an evolutionary fashion, and there will be that middle ground for a while,” he says. “What I’m saying is, let’s keep that middle ground to as short a time period as possible.”
The self-driving marketplace is going through a preliminary growth spurt, as several auto manufactures have stated plans to bring models to market over the coming years. GM announced large investment in the technology in early June, while BMW plans to bring a fully autonomous vehicle to market, dubbed iNext, by 2021.
Related Links:
Deadly crash raising self-driving car safety questions
GM to enter self-driving car market
