Somehow the creators of "The Jetsons" never got around to sharing with us exactly how those self-driving -- not to mention self-flying -- cars of the future will be insured. (They never addressed exactly how George Jetson's flying pod transformed itself into a briefcase upon landing either, but I'll leave that for Tara Baukus Mello over at Bankrate's cars blog to figure out.)
Auto insurance actuaries are finally confronting the Jetsons' coverage question, and not a moment too soon. Google says it will introduce a fully-automated vehicle by 2017, and major automakers predict their versions will roll themselves onto the red carpet by 2020.
Given the fairly compressed timeline, top actuaries brainstormed the underwriting road ahead during the "Autonomous Vehicles and the Impact on the Insurance Industry" symposium, held recently at the Ratemaking and Product Management Seminar of the Casualty Actuarial Society.
New questions for underwriters
If you've been wondering why auto insurers have been revving up their data-collecting in recent years, the answer runs far deeper than merely generating marketing ideas for Flo, Mayhem and the Geico gecko. After a century of figuring out -- and underwriting for -- what are now the fairly predictable risks of the American road, the prospect of doing likewise when cars commence to self-brake and navigate by themselves is akin to starting from scratch.
On the plus side, most agree that turning the wheel over to, well, the wheel will greatly reduce auto accidents, given that human error is to blame in more than 90 percent of collisions. A Princeton study concluded that turning the keys over to the car could save you $475 a year on auto insurance.
But to get there, insurance actuaries will need to gather crazy amounts of data on things that scarcely entered the underwriting frame before.
Lots of new variables
For starters, each self-driving car is expected to generate 750 megabytes of data per minute. Actuaries will somehow have to sort through that digital mishmash in order to identify and analyze the predictive variables, the rudimentary building blocks of determining risk and setting auto insurance rates.
Individual auto brands will likely pose another challenge, since the variables involved in predicting a BMW crash will probably differ from those of a self-driving Hyundai.
Then there's the changing traffic mix. In the short term, there will be an ever-evolving mix of three kinds of vehicles on the road:
- Fully autonomous, in which the driver does nothing but start the vehicle.
- Partly automated, in which the driver does some or most of the driving.
Robots and no-bots
It's that lengthy and actuarially confounding stage between robot and no-bot during which driving functions shift en route between human and software that's likely to challenge drivers and underwriters the most.
"How do you keep somebody sharp enough to take over the car?" asks Frank Douma, a research fellow at the University of Minnesota.
As self-driving vehicles become the norm, vehicle liability will likely change seats from the driver to the vehicle manufacturer and/or the company that installed the automation systems.
By comparison, cramming a car into a brief case doesn't seem so difficult.
Here's more on the fast-approaching dawn of self-driving cars.
Follow me on Twitter: @omnisaurus.
Get more news, money-saving tips and expert advice by signing up for a free Bankrate newsletter.