New Technology Requires New Approaches
The future is here, and self-driving cars are on the road in some cities. Waymo is currently operating a fleet of fully autonomous vehicles in Phoenix, San Francisco, Los Angeles, Austin, Atlanta, and Miami, with greater expansion to come in 2026.[1]
Meanwhile, automakers are making progress toward passenger vehicles with Level4 (L4) autonomous driving capabilities being available. L4 systems are designed to respond to their surroundings without the need for human drivers.[2]
And yet, while the technology is advancing, there are plenty of legal and insurance complexities to work out. While some states have laws around self-driving cars, there are no federal laws yet.[3]
Self-driving cars sit in an awkward transitional moment for the insurance industry. The technology is moving quickly, but insurance frameworks were built on the assumption that a human driver controls the vehicle. As automation increases, that assumption becomes less reliable and harder to insure with confidence.
Fully autonomous vehicles challenge that structure in a more fundamental way. When no human is expected to monitor or intervene, responsibility begins to shift away from individual behavior. The focus moves toward system performance, software reliability and operational oversight. Insurance must adapt to a world where the “driver” may be code.
This shift changes how auto insurance thinks about risk. Traditional personal auto underwriting centers on who you are, where you drive and your prior driving history. Autonomous technology reduces the relevance of those factors and elevates technical reliability and system design. Risk becomes less personal and more institutional.
Partially automated vehicles create a shared-control problem. The system may handle routine driving, but the human is still expected to step in during unexpected situations. When a crash occurs, insurers must decide whether the driver failed to intervene or the system failed to respond appropriately. That gray area complicates liability decisions.
Why Claims and Liability Get Messy With Automation
Claims handling looks very different when automation is involved. Here are some of the key differences:
1. Evidence and documentation changes.
Adjusters must evaluate digital evidence such as event data recorders, sensor inputs and system engagement logs. Traditional statements and police reports are no longer sufficient on their own. The claim becomes as much a technical review as a factual one.
2. Claims resolution may slow down.
These additional layers tend to slow down claims resolution. Multiple parties may be involved, including vehicle owners, manufacturers, software vendors and fleet operators. Each has a financial incentive to frame the data in a way that limits exposure. That can prolong disputes and increase loss adjustment costs.
3. Subrogation matters.
Subrogation plays a larger role in self-driving claims. An insurer may pay a claim under a personal or commercial auto policy and then seek recovery from a manufacturer or operator. These disputes increasingly resemble product liability cases rather than routine auto recoveries. Legal complexity becomes part of the loss cost.
How Insurers Are Adapting—and Why It’s Still Complicated
The commercial autonomous vehicle model highlights these issues even more clearly. Robotaxi and delivery fleets typically carry high liability limits to account for worst-case scenarios. Even with lower crash frequency, the severity of a single incident can be significant. Insurance pricing must reflect that imbalance.
1. Underwriting is looking beyond just the individual vehicle risk.
Underwriting these fleets requires a broader assessment of risk. Insurers evaluate not just the vehicle, but the entire operational framework supporting it. Maintenance schedules, software update controls, remote assistance protocols and safety monitoring all factor into pricing decisions. The vehicle itself becomes only one piece of the exposure.
2. New pricing models.
Some insurers are experimenting with pricing models that separate human-driven miles from automated miles. This treats automation usage as a dynamic underwriting variable rather than a static feature. It reflects the belief that risk changes depending on who—or what—is in control. This approach marks a meaningful departure from traditional rating practices.
As long as the technology within the auto industry keeps improving and getting optimized with continuous testing, it won't matter much whether you embrace the self-driving possibilities now or not. Regardless, know that they might be here to stay for good. Of course, self-driven cars will change how we drive, but it's too soon to determine the extent of the changes.
Wrapping Up
Self-driving car insurance is ultimately less about eliminating risk and more about relocating it. Responsibility is moving away from individual drivers and toward systems, operators and the organizations that design and maintain them. Until the legal, regulatory, and technical questions are settled, insurers will continue pricing not just accidents, but uncertainty itself.