Risk Assessment of Automated Vehicles: House Hearing on the Future of Insurance and AVs

May 24, 2018

Today, the safety risks associated with operating motor vehicles are so high that drivers are required to obtain some level of liability insurance in most states. Significant market penetration by automated vehicles (AVs) has the potential to transform that insurance market. Theoretically, safety on the roads could be increased to a point that people can move in AVs without assuming any notable amount of risk, the car insurance sector will lose that business. (This scenario is not yet a plausible reality, as the cost of AVs and the death of Elaine Herzburg and other crashes have proved.) And practically, some liability for crashes might be transferred from the human driver to the (much deeper-pocketed) company that manufactured and sold the vehicle.

The challenge is to figure out how to assess risk with AVs and other emerging technologies. Panelists representing the insurance sector at Wednesday’s House Committee on Financial Services hearing “The Impact of Autonomous Vehicles on the Future of Insurance” emphasized the need for insurance companies to have access to data from AVs to be able to adequately assess risk.

Witnesses from State Farm and American Family Mutual Insurance Company noted their collaboration with research groups at universities such as testing projects at Mcity. Access to data from testing groups and researchers is useful, but more robust data sets would improve analytic capabilities. Beyond risk assessment based on the technology functionality, companies will need data on the performance of vehicles in real world transportation systems. Depending on how the models work, aggregated data sets that do not reveal personally identifiable data or company trade secrets may be able to assist insurance companies in risk assessment.

Today, most vehicles do not track personal information such as eye movement to identify distraction or even acceleration profiles to identify reckless driving behavior, but existing tools such as Georgia Tech’s second-by-second Commute Warrior tracking app can provide these data to people and to their insurance companies. The technology exists, but personal risk assessment utilizing it has not been widely adopted—a trend that may very well continue as vehicles grow to collect more and more data about users, unless a policy intervention begins to require certain data collection and sharing practices.

The insurance industry witnesses at the hearing suggested that AVs are inherently different enough from today’s vehicles that the model could even completely shift from an insurance perspective to a product liability perspective.  The functionality of this option was not further explored, and the future of AV product liability most likely lies in case law that is yet to be established. A complete overhaul of the insurance model for vehicles brings up some difficult questions though, such as how to define “autonomous vehicle.”

Other shifts in current models could be generated by changes towards mobility as a service, and ownership models. In anticipation of shared fleet models of vehicle ownership, insurance companies are exploring possibilities to change their models to personal mobility insurance instead of vehicle insurance.

All of these shifts in insurance models would likely have to be legislated at the state level, but federal requirements for data sharing and policies that could have an impact on ownership models can greatly influence how the insurance industry reacts to the emergence of AVs.  Witnesses impressed upon Congress its role in crafting policies such as the impending AV Start bill to clarify data rights.

Witnesses at the hearing included:

Search Eno Transportation Weekly

Latest Issues

Happening on the Hill