How We’re Encouraging Safer Autonomy Through Insurance
Introducing Edge Case Risk Management
By Benjamin Lewis and Michael Wagner
Safety is the most important hurdle that autonomous vehicle technology faces as we prepare for its deployment on our highways and into our neighborhoods. Polls consistently show the public is concerned about the safety of autonomous vehicles despite recent progress across the industry. Just last year, AAA and Harvard Kennedy School’s Belfer Center for Science and International Affairs found that, among surveyed U.S. drivers, 36% opposed autonomous vehicle testing on public roads, with safety cited as their top concern. Clearly, the issue of safety must be a top priority for the industry if acceptance and adoption of autonomous vehicles are to be achieved.
For autonomy to deliver on promised benefits, the industry must validate that its technology is safer than human drivers, as the losses on our roads today are significant and many crashes have preventable causes. These losses have continued to increase, even as we have all driven less amid the pandemic, with NHTSA estimating that 31,720 people died in motor vehicle crashes from January through September 2021, an increase of approximately 12% over the same period in 2020. That is the highest percentage increase during the first nine months of any year since NHTSA implemented its current reporting system in 1975. While it is very likely autonomous vehicles will do better and address many of those preventable causes, meeting this residual risk threshold with such novel technology at scale is an enormous challenge.
Two years ago, Edge Case Research partnered with UL Standards and a panel of industry leaders to define, for the first time, how to validate the safety of autonomy. The UL 4600 standard is the result of that effort. UL 4600 lays out a long list of expectations in its three hundred pages, but what it doesn’t provide is step-by-step instructions for exactly how to meet those expectations.
Six years before UL 4600 was published, Edge Case was founded to define those instructions and build the evidence-based safety arguments autonomous vehicle technology developers need in order to answer the question “is it safe enough to deploy?” Since 2014, our team has pioneered techniques and technologies that assess the safety of autonomy in smarter and more scalable ways. More recently, we built nLoop so that members of the industry can measure, manage, and communicate the safety of their autonomous vehicle technology as it is developed, tested, and operated on our roadways. At the same time, leading players in the autonomy industry have been standing up top-notch teams to manage safety processes across their organizations. This is fantastic progress. However, significant investments are still needed to meet the industry’s risk objectives.
Insurance can drive smart investments in safety
Autonomy promises to reduce fleet operating costs, of which insurance is a major component. Today, however, insurance carriers cannot price premiums in a way that accurately and specifically reflects the risk profiles of autonomous fleets. Software drivers and their creators are evaluated by insurers in much the same way as traditional vehicle risks because the complex engineering processes and technology behind autonomous vehicles must be met with equivalent technical expertise that lies outside of the insurance industry itself.
In short: how can insurers and reinsurers properly weigh and price the risk of systems for which there is a lack of historical data and in-house expertise to evaluate what data there is? How can brokers advocate for better rates on behalf of their clients without a compelling, evidence-based argument and an intelligible framework to communicate those claims to underwriters?
Realistic, technical, data-driven assessments of autonomous vehicle risk will enable the insurance industry to price risk accurately, benefitting insurance buyers as well. This is where we see opportunity. We intend to build an insurance market that offers actionable insights to risk managers and rewards real reductions in residual risk with lower premiums. Such a market will incentivize and reward developers and operators who invest in safety engineering.
Some may ask why we, a company steeped in engineering and technology, have decided to embark on this journey to build a business centered on insurance for autonomous vehicles.
One reason is that, after safety, liability is top of mind for the public and policymakers when it comes to autonomous vehicles. The 2021 AAA and Harvard survey found that liability determination in the event of a crash was a close second among the concerns of those opposed to autonomous vehicle testing on public roads. At Edge Case, we believe the two issues are interconnected and our techniques and technologies are tailor-made to bring clarity to both.
Another reason lies in our experience. Over the past four years, Edge Case has partnered with some of the world’s leading insurers and reinsurers to help them meet the complex insurance and risk management needs of the emerging autonomous vehicle ecosystem. As Edge Case expanded our business with insurance companies, we identified three areas where we could add value:
- Proactive mitigation during the development process as a key tool to manage the risks posed by modern software-enabled products. Historical actuarial data remain significant, but the residual risk of an autonomous vehicle could shift constantly as new versions of software are delivered over the air. Over reliance on historical data will delay the recognition of evolving risk and inhibit insurers’ ability to offer rewards for mitigation.
- The shift to highly correlated risks stemming from software drivers. Insurers’ actuaries can reliably model claims rates from a huge population of independent human drivers. But carriers will now need to assess and price systemic risks that span entire fleets of vehicles equipped with the same software driver. The same methods applied to price driving risk today will not be sufficient in the world of widespread autonomous mobility.
- Speed to market. Autonomous vehicles are driving on our roads today, and operations will scale dramatically in the next few years. Yet insurance products have not kept pace. Now is the time for the industry to make and sustain substantial investments in safety, and for insurers to similarly invest in understanding the risk the technology presents.
The remaining question is: how can we build such a market?
Ever seeking solutions, and in support of safer autonomy, Edge Case is announcing a major step to meet these needs. In partnership with primary insurance carriers, reinsurers, and brokers, we are establishing Edge Case Risk Management, a managing general agent (MGA) that will be dedicated to underwriting autonomous vehicle developers and their customers.
We will begin by offering insurance products that tie directly to companies’ safety cases and the associated metrics. Our business will scale with our customers as they go to market. While we expect that autonomous trucking fleets will be the first large-scale deployments, we will also expand into ride-hailing, goods delivery, and other segments where autonomy is likely to be adopted at scale. In the US, the market for auto liability insurance alone generates well over $300b in gross premiums annually, and it is expected that a significant portion will be displaced as autonomous vehicles become prevalent on our roads. With our partners, we aim to take a leading position as this market solidifies and be there for customers as they commercialize in the coming years.
Edge Case Risk Management underwriting will leverage the expertise of the global Edge Case team, which has more than a century of combined experience assessing and developing safety arguments for autonomous systems. Prior to joining the company, members of our talented team helped build simulation infrastructures for Uber ATG and Argo AI, advised on safety at Aurora, tested combat systems at the U.S. Army Aberdeen Proving Ground, and researched the implications of safety-critical AI at Carnegie Mellon University. Building on that, the Edge Case team has successfully helped four major automotive OEMs, a half-dozen software driver developers, four Tier 1 suppliers, and a number of other market participants to make their systems safer and more reliable. Our team of system safety engineers also outnumbers even large autonomous vehicle developers.
All of this means that Edge Case has the scale, the software, and the experience to be able holistically assess the risk of any autonomous system, to evaluate the validity of a safety argument, and to help our customers understand and pursue what thresholds they need to meet to make deployment decisions based on safety.
We will continue to provide that valuable safety engineering support and software alongside the insurance coverage that customers will be able to access via Edge Case Risk Management; all in order to deliver the promise of autonomy.
We hope you will join us on this journey to recognize and encourage safer autonomy. We have begun building a network of insurance brokers, carriers, actuaries, and technology providers. If you are committed to building autonomy safely from the ground up, we’d love to work with you, too.
About Edge Case Research:
At Edge Case Research, we believe that every autonomous vehicle should be built safely from the ground up. We help companies go to market with autonomous mobility that is safe, reliable, and trustworthy. Founded in 2014 by leaders in autonomous vehicle safety, our expert team provides system & safety engineering services, nLoop Live Safety Case software, and risk management solutions to software driver developers, automotive and trucking OEMs, Tier 1 suppliers, insurance providers, and the aerospace and defense industries. For more information, visit ecr.ai or ecr-defense.ai.