For those who think life would be easier with a chauffeur to quicken the commute, technology may soon provide a way to Easy Street. Testing of autonomous vehicles, also commonly referred to as driverless or automated vehicles, is showing rapid advances in reliability and safety, pointing toward a day when hands-free transportation could make its way into the fast lane.

How will insurers react? And who's responsible when a driverless vehicle goes awry?

As soon as 2020, semiautonomous vehicles could become more broadly available, according to several car manufacturers. At least one manufacturer suggests it may bring a fully autonomous vehicle to market as soon as 2017. These predictions change quickly as manufacturers are increasingly upping their efforts on this front. But even if these predictions come to pass, it will take time before autonomous vehicles become commonplace.

The first driverless cars are likely to be expensive and may well be out of reach for most consumers. The novelty and potential safety risks of the technology may further limit its use to a still smaller group. These factors, along with increases in the useful lifetime of an average car, cause some to estimate that widespread adoption of autonomous vehicles may not become reality until 2040, 2050, or even 2070. Still others predict that obstacles facing a truly practical autonomous vehicle may be insurmountable, although this opinion appears to be in the minority.

Are driverless cars safe?

Recent vehicle safety and technology improvements include the rearview camera, automatic braking, adaptive cruise control, and lane-control warnings. Other driver-assist features like forward collision warnings have also advanced over the last few years. The National Highway Traffic Safety Administration classifies vehicle automation in five levels, ranging from Level 0 (no automation) to Level 4 (full self-driving operation, with the driver "not expected to be available for control at any time during the trip"). As shown by recent developments, the state of vehicle autonomy is well on the way to Level 4.

Many companies, including technology firms and traditional auto manufacturers, have already started to deploy semiautonomous and handsfree vehicles in a few jurisdictions, as well as in controlled environments such as Mcity, an artificial metropolis designed by the University of Michigan's Mobility Transformation Center in Ann Arbor. These sites test automated and connected-vehicle technology using a wide variety of simulated conditions. While testing may help to refine the technology and establish that it works, large questions remain about where and when autonomous vehicles will be ready for use on public roads and highways.

Determining liability

So where is the autonomous car in 2015? And what are the insurance implications for self-driving vehicles? A few perspectives must be considered when answering the question of how autonomous vehicles will affect the insurance industry. One of the issues now being debated is where could the fault, or liability, lie when an autonomous vehicle crashes? Would the owner of the vehicle be held responsible? Will the liability shift to the manufacturer of the vehicle or safety feature?

Today, when speaking of liability exposures, insurance is largely focused on the driver. This holds true in terms of a number of state legislative requirements (with respect to financial responsibility and/or compulsory liability insurance) as well as in personal auto rating plans. From an insurance standpoint, the focus has largely been on the assessment of which driver is liable when one car hits another car—whether the cause is speed, lack of attention, or some other factor. Insurance rating mechanisms typically take into account various driver attributes, including a driver's history of accidents and violations, or lack of them.

When looking at liability and collision risk, the model or technology of the vehicle itself is only one aspect that various rating plans take into account. Yet a number of states have regulated auto insurance by generally removing the aspect of fault for injury. In these "no-fault" states, the idea is essentially to eliminate the matter of determining liability in most accidents.

With regard to loss experience, it's almost universally accepted that autonomous vehicles will likely reduce automobile insurance losses to some extent. The industry is already seeing improvements from advances in safety features built into driver-assist systems, which have brought reductions in accident frequency and bodily injury severity and should continue to do so as they become more widespread. But these reductions are somewhat offset by the increased cost of repairing the new technology.

Over time, autonomous vehicles will likely bring about further loss reductions because processors are capable of reacting more quickly to sensory input than humans can and won't become distracted or fatigued. The technology is also expected to become less expensive, and repair costs will be less of a concern.

Effects on coverage

How insurance coverage itself will be affected is less certain. Automobile insurance today is generally well understood, and questions about liability are likewise relatively easy to address. But the question remains: When an autonomous vehicle is involved in an accident, who is at fault? Some have suggested that one potentially less disruptive change to auto insurance could be to move toward a no-fault system to avoid this stubborn question entirely. Such a reform could also help to address the mixture of vehicles on the road operating under varying levels of autonomy.

Alternatively, product liability insurance may one day completely replace auto liability in cases where driverless vehicles are determined to be the cause of an accident. The personal auto insurance market may also be expected to shrink because of other developments already under way, such as the growth of the sharing economy and commercial auto insurance taking on more exposure as a result. There may come a time when personal auto insurance is considered a specialty line of business and relegated to covering "antique" nonautonomous vehicles.

Even so, auto insurers can probably bank on at least one source of premium. Until the next phase of technology comes along to reduce vehicle thefts, either thefts in person or someday perhaps remotely through cyber hacking, there will still likely be a need for theft coverage.

The authors

Sandee Perfetto is director of personal auto/umbrella product development, and Jared Smollik is the actuarial product director for personal auto at ISO, a Verisk Analytics (Nasdaq: VRSK) business.