What is an autonomous vehicle? As demonstrated in a Sept. 15 California Department of Insurance informational hearing, that’s a misleadingly simple question.

Regulators’ attempt to stay in front of the curve, by delving into how autonomous vehicles will interact with insurance, is sensible in light of the regulatory obstacles presented by Prop 103. The emergence of this new suite of technologies actually belies a complicated narrative.

In truth, the era of semi-autonomous vehicles is not waiting to arrive; it is here now. For instance, collision avoidance systems, such as those warning of unintentional lane departure, while seemingly minor in their interference, place a vehicle on the autonomous vehicle spectrum.

The National Highway Traffic Safety Administration categorizes levels of vehicle autonomy into five tiers, from zero to four. Level zero envisions no automation whatsoever. “I think of it as a ’55 Chevy,” Princeton University professor Alan Kornhauser told the CDI. On the other end of the spectrum, at level four, is a vehicle that neither requires nor accepts driver input for critical driving functions. A level four vehicle would have no steering wheel or pedals. Today’s new vehicles largely function at level one.

Where it gets interesting for questions of insurance coverage is at the intersection of levels two and three. At level two, a vehicle may have control over a number of major functions, but still require constant driver attention. Level three sees the need for driver attention reduced to the extent that drivers can focus on other activities. Between these two levels, delineating fault and liability in the event of a collision moves from difficult to, perhaps, impossible.

Since driver input at levels two and three is not constant, evaluating a collision to determine when a driver is in control – or in the process of continually regaining and relinquishing control, and thus responsible for driving – is a labor-intensive proposition. Another layer of complexity is introduced by the prospect of level two and three vehicles communicating with one another to coordinate their activities. In that case, determining which vehicle was the genesis of a causal, collision-inducing action may prove impossible.

Loss scenarios that complicate insurance considerations should not forestall rapid autonomous vehicle development; instead they should inform early discussions about rating autonomous vehicles. The CDI requires insurers to base their rates on actuarially sound data. This is, of course, impossible when there is no data, but instead, only the prospect of favorable outcomes. Insurers are left in a bind.

Since nothing ever works perfectly, not all autonomous safety augmentations will achieve all of the risk reduction for which they are designed. One manufacturer’s system might outperform another, while another manufacturer’s system might promote or augment operator behavior in a dangerous way.

Without loss data, the ability of actuaries to promulgate an accurate picture of the relative risks of level two and three autonomous vehicles will be hamstrung. Insurers would not be able to immediately offer rates that reflect the promised risk-reduction of autonomous-vehicle technology. Here, the CDI can play a positive role by embracing rating flexibility.

It is indeed encouraging to see the CDI hold a meeting to anticipate the new world of automobile insurance in a robotic age. The hearing solicited testimony from academics, stakeholders and the public. Participants drew attention to an array of pitfalls that the state could encounter or foster if it is indelicate in its approach.

Hopefully, continuing efforts will be focused directly on solving the inevitable problems posed by today’s gradually less-relevant, and increasingly burdensome, insurance laws. Perhaps the CDI could induce insurers to take the lead in laying bare anticipated problems and providing timely solutions. The CDI has the power to stimulate insurers’ problem-solving prowess, and it should, since it has neither the profit motives nor practical expertise to do the job itself. It will need to assist insurers by advocating, in legislation and regulation, changes to modernize insurance law appropriately before relentless change overtakes both it and California drivers.