Henry Kowal, Arity's director of insurance product, shared with Digital Insurance about new car technology and how location is important data for self-driving risks.
What risks are associated with new car tech?
Some drivers may overestimate what these assistance features can handle and pay less attention on the road. This gap between perception and reality can increase risk, and there have been reported crashes involving driver assistance systems that highlight the stakes.
Another risk factor is that many safety and driver-assisted features can be disabled. Drivers sometimes turn them off if alerts are distracting or feel intrusive. When that happens, a vehicle may be equipped with safety tech but not actually benefit from it — which matters for real-world risk and how insurers understand exposure.
Finally, drivers should also consider system complexity when looking to invest in new smart cars. As vehicles become software-reliant, that software becomes part of the risk equation. It's important to remember that connected car technology can improve safety, yes, but they also expand the potential exposure to digital risks.
What is different about self-driving risk in some locations?
Self-driving risk is rarely uniform because location varies in terms of traffic density, road design, weather, and even driving culture. Those differences can be significant, with different risk outcomes from one region to another.
Arity's data has
That's why context matters so much. An AV system trained primarily in sunny, low-density environments may face very different risk cases in dense, mixed-use urban cores or regions with snow and heavy rain. Understanding these geographic nuances is essential for insurers, policymakers, and AV developers alike when understanding the full picture of risk and a realistic view of safety performance.
Why should insurance companies be thinking about the various risks of AV across different scenarios and locations?
The risk landscape is evolving fast and becoming more complex as AVs hit the roads. The core challenge is that conventional actuarial frameworks were built on loss history — and for AVs, that history is still thin and not very credible.
That's where behavioral and mobility data become especially valuable. They give carriers a forward-looking granular view into how risk is forming and shifting across different geographies. A dense urban area behaves differently than a suburban one. Weather, traffic patterns, and local driving culture all factor in. Insurers need to be thinking at that level of granularity because AVs will be making deployment decisions at that level, and coverage, underwriting, and pricing decisions have to keep pace with how the real world is changing.
Territorial ratemaking is one of the clearest applications of this approach. Rather than relying solely on lagging indicators such as historical loss data, carriers can use real driving behavior data at granular geographic levels to more precisely assess risk and AV exposure by location and adjust as conditions evolve.
The insurers building that data infrastructure now will be significantly better positioned as the market scales.
How can mobility data show where safety incidents happen?
For insurers, the challenge has always been getting ahead of risk rather than reacting to it. Mobility data offers a way to do that — surfacing leading indicators like hard braking, speeding, phone distraction, and changing traffic patterns that reveal where and when risk is building rather than relying on lagging and backwards looking claims data.
This shifts how carriers can think about underwriting, moving from "where have losses happened?" to "where is risk accumulating right now?" Rather than pricing off of historical claims alone, insurers can incorporate real behavioral signals into how they assess risk.
Mobility data can also improve AV safety by identifying locations and conditions that create more complex driving environments. Areas with risky human driving behaviors and conditions — such as intersections, road segments, and time of day — can inform AV testing, validation, and safety measures.
The state of road risk is volatile right now and will continue to evolve as more AVs enter the mix and human drivers adapt around them. Mobility data will give carriers a way to track that shift, so their pricing and coverage decisions stay calibrated as the landscape changes.
Anything else you would like to share with me?
The takeaway isn't that new tech is inherently risky. It's that risk is evolving. We need better data to understand how humans interact with these systems in real-world conditions, not just controlled environments.
We're in a transition period. Advanced driver systems and AV pilots are expanding, but human behavior remains central to road safety. AV safety will ultimately be shaped not only by how an automated system performs, but also by how it coexists with the human road ecosystem around it.
As consumer sentiment evolves alongside the technology, there is an opportunity to use mobility data responsibly to understand interactions between humans and machines, in real geographies and real-world conditions. Insurers, automakers, and cities should approach AV pilots more collaboratively and use mobility data to align innovation with safety priorities as AV technology scales.
