The Drive Towards Fully Automated Vehicles Continues
When a Tesla vehicle with its Autopilot system engaged was involved in a fatal accident in Florida on May 7, it was the first one involving a partially automated Level 2 driving system. Since then, this accident has been the subject of scores of articles, including a blog by my colleague Mathias Schubert on the accident's specifics.
The sentiment has ranged from This will force a slowdown in the introduction of driverless technology, to While any fatal accident is a tragedy, this will not slow the inevitable quick march towards driverless vehicles.
Let's not forget, however, that the primary reason for moving toward fully automated vehicles is contrary to what many of us believe; we're not very good at driving. According to the National Safety Council, there were some 38,300 fatal auto accidents in the U.S. in 2015 as well as 4.4 million people injured.1 Fully automated vehicles represent the promise of a dramatic reduction in both frequency and severity of accidents. Predictions range as high as an 80% to 90 % reduction. Even if one believes that such lofty predictions are unrealistic, a recent article in Insurance Thought Leadership stated, “A 25% reduction in auto-accident-induced fatalities would save more lives than curing leukemia; a 75% reduction would save more lives than eliminating suicide.” Pretty compelling reasons not to slow down such technological advancements.
Obviously, we are not there yet; we have seen additional reports of accidents possibly involving Tesla's Autopilot. These include an August 10 report of a Tesla vehicle in China with its Autopilot engaged hitting a car "parked half off the road;" on August 19 a Tesla Model S reportedly in Autopilot mode allegedly steered off the road into a guardrail in Kaufman, Texas. The driver admitted he was not paying attention due to his trust in the Autopilot system. The driver sustained only minor injuries but the car may have been totaled.
Despite those accident reports, automakers and tech companies from around the globe continue to invest in driverless technology and seemingly race towards fully automated vehicle production. For example, here are just a few announcements since the May 7 fatality:
- BMW will have a "fully driverless car" on the market by 2021.
- Jaguar Land Rover will begin testing driverless cars in the UK in 2016.
- Audi will launch self-driving car in 2018 to handle up to 80% of driving situations.
- GM will use Lyft to launch its first self-driving car. Mercedes-Benz launched its self-driving, full-sized bus.
- Auto parts suppliers Delphi and Mobileye team up to develop a self-drive system that car makers could begin placing in their vehicles beginning in 2019.
- Highways around Columbus, Ohio will begin testing driverless platooning tractor trailers starting in 2016.
In addition, the DOT and NHTSA (National Highway Transportation Safety Administration) are also strongly pursuing "new life saving technologies…and tools that save lives."2
The current technology used in Tesla's Autopilot system and those of other manufacturers working on partially and fully automated vehicles will certainly improve. Tesla itself will soon introduce the next generation of the Autopilot system, bringing it to Level 3 automated driving, which the NHTSA defines as:
"Limited Self-Driving Automation (Level 3): Vehicles at this level of automation enable the driver to cede full control of all safety-critical functions under certain traffic or environmental conditions and in those conditions to rely heavily on the vehicle to monitor for changes in those conditions requiring transition back to driver control. The driver is expected to be available for occasional control, but with sufficiently comfortable transition time. The Google car is an example of limited self-driving automation."3
Additional accidents will, inevitably, occur. The key issue is when will they become - or have they already become - better drivers than we are? According to a few news reports, Teslas on the road had already accumulated 130 million miles of driving while on Autopilot by the time the fatality occurred. Compare that to about 94 million to 100 million miles driven by humans per fatality. Some authors argue that such systems are already better than human drivers. Others have stated that the above comparison isn't valid. On the other hand, given the track record of human driving, if such systems are not yet better than we are, they soon may be.
Insurers have some time to come to grips with this potentially disruptive technology, though, as one can see from the most recent announcements listed above, maybe not as much time as some have thought.
- "2015 Brought Biggest Percent Increase in U.S. Traffic Deaths in 50 Years," Newsweek, 2/17/16.
- "Tesla crash won’t cool federal enthusiasm for automated cars, top safety official says," The Washington Post, July 20, 2016.
- "U.S. Department of Transportation Releases Policy on Automated Vehicle Development", NHTSA, May 30, 2013