The crash of a vehicle operating in semi-autonomous or fully autonomous mode presents a headline-grabbing opportunity to question the technology and the pace at which it is being introduced. Every accident resulting in injury or death is a tragedy. In the case of new technology that offers the possibility of dramatically reducing the total number of injuries and deaths, it will be important to look at any individual crash in the context of the overall promise of the technology. For example, it will be important to know how many miles of autonomous driving took place before the first crash occurred and compare that to ordinary cars, where the national average is one fatality every 94 million miles, and the worldwide average is a fatality every 60 million miles.
For our purposes, a crash also presents a scenario for how liability and insurance issues may play out as these cars and trucks start appearing on the road in greater numbers. We’ll walk through what won’t change, and the few things that might.
First, the car manufacturer, the driver, and any other car or truck involved will still have to sort out the cause of the crash and liability for it. The process for doing this (with one key exception) and the product liability and negligence law applicable to it will follow those of any other car crash. Product liability law will come into play in the same way as if an ordinary brake mechanism may have failed. The parties will still need to determine whether the autonomous technology failed in some way, or whether the manufacturer did not sufficiently warn of weaknesses in the system. If it failed, was it software or hardware, and who created the flawed technology? The driver may bear responsibility to the extent that she did not follow instructions for how to use it safely (i.e., keeping her hands on the wheel and eyes on the road). It may also be the case that the driver of the other car or truck was driving negligently and created the hazardous situation, to which the autonomous technology did not respond well enough to prevent the crash.
The one key exception for sorting this out is that the National Highway Transportation Safety Administration (NHTSA) may intervene because of its keen interest in autonomous car technology. For these accidents in the early adoption phase of the technology, the costs of determining what happened and why may fall largely on taxpayers, not on the litigants. It will be worth watching whether the NHTSA feels compelled, in the short run anyway, to investigate virtually all accidents involving autonomous car technology the way the NTSB investigates aircraft accidents.
Moreover, insurance will apply to the crash as in any accident involving an ordinary car. Auto insurance for owners of cars even with the most advanced autonomous driving features is no different than for any other car (except perhaps for the premium). Insurers have not attempted yet to add language trying to address the use of autonomous driving features like advanced cruise control. They seem to be in a wait-and-see posture while data accumulates to show how safe these features are.
We have run across only one policy so far that bills itself as a “Driverless Car” policy. Written by Trinity Lane Insurance Company (a Malta-based insurer) for the British market, it expressly agrees to cover the driver if the autonomous systems fail, or if the owner failed to install updates to the system software in a timely manner. But in Trinity Lane’s standard policy (and in every other auto policy we are aware of) these causes of an accident are covered anyway. It turns out that the purpose of this language in the Trinity Lane policy is to add an additional £250 deductible to the claim if the accident involved the owner not timely installing a system update. (By the way, it’s not surprising that a Driverless Car policy should first appear in Great Britain, where the government has openly supported the testing of autonomous car technology on public roads in a bid to give British companies a competitive advantage.)
While incremental and somewhat awkward, Trinity Lane’s policy does point to one approach insurers could take. While not limiting coverage for autonomous cars or features, they could offer a mix of carrots and sticks (lower premiums, higher deductibles, a more robust application process) to limit their risks. Insurers might offer lower premiums to owners who have received some training in how to drive a car with autonomous features, in particular, how not to zone out while the system is engaged and how to take control if it fails. We see this approach now with another new technology — commercial drones – where insurers may offer lower premiums to operators with greater training and experience. They could, for example, like Trinity Lane, demand a higher deductible if it were shown that the driver did not have her hands on the wheel. The goal would not be to disclaim coverage if an accident occurred involving autonomous technology, but to screen and incentivize their insureds to pay attention despite the technology.
If virtually every major car manufacturer and the NHTSA are right, the introduction of autonomous car technology will reduce the number and severity of accidents, lowering premiums for everyone. Insurance for drivers will remain essential, and policy wording might even remain unchanged. We will, though, more likely see insurers watching the data closely and innovating themselves to better respond to the rapidly changing world.