Who pays when autonomous vehicles cause accidents determining legal liability

Unlike traditional car accidents where the driver is almost always liable, autonomous vehicle accidents distribute liability among manufacturers, vehicle...

When a self-driving car crashes into another vehicle or pedestrian, the question of who pays damages isn’t answered by a simple rule—it depends on whether the vehicle’s software, hardware, the human owner, or multiple parties share responsibility. Unlike traditional car accidents where the driver is almost always liable, autonomous vehicle accidents distribute liability among manufacturers, vehicle owners, insurance companies, and sometimes government entities, based on what actually caused the collision. If a sensor malfunction or software defect caused the accident, the manufacturer may be liable; if the vehicle owner failed to maintain the car or disabled safety systems, they might be responsible; if a human driver was supposed to be monitoring the vehicle and wasn’t, liability could fall on them instead.

This liability landscape is still evolving because most states lack specific autonomous vehicle laws, and only a handful of accident cases have been resolved in court. The insurance industry has begun to shift how it handles autonomous vehicles, some manufacturers have purchased their own liability insurance for their vehicles, and federal regulators are still developing standards. For anyone injured in an autonomous vehicle accident, understanding who can actually pay a claim is crucial—because filing suit against the wrong party means your claim might be denied, delayed, or underfunded.

Table of Contents

liability in autonomous vehicle accidents typically falls into four categories: the vehicle manufacturer, the vehicle owner or operator, a third-party software or component developer, and in some cases, the government entity that maintains the road. Manufacturers can be held liable if their vehicle design, software, sensors, or safety systems were defective or failed to function as promised—this includes whether the manufacturer adequately warned users about the system’s limitations. When Tesla’s Autopilot was involved in accidents where drivers assumed it would handle all driving, investigations revealed that Tesla marketed the feature as more autonomous than it actually was, which shaped liability arguments in subsequent lawsuits.

Vehicle owners face liability if they failed to perform required maintenance, disabled safety features, or knowingly allowed someone to operate the vehicle improperly—this is similar to traditional negligent entrustment claims. Third-party developers who provided software or sensors to a manufacturer might also be liable if their component was defective. Government agencies can potentially be held liable for poorly maintained roads, missing signage, or road conditions that autonomous vehicles failed to navigate safely, though sovereign immunity often protects government defendants. The key question courts ask is: who had the duty to prevent this accident and who breached that duty? In a 2016 Tesla Autopilot collision that resulted in a fatality, initial investigations suggested the vehicle’s autopilot might have contributed, which immediately raised questions about whether Tesla bore responsibility despite the driver being behind the wheel.

Manufacturer Liability and Product Defect Claims

Vehicle manufacturers face the strongest liability exposure in autonomous vehicle accidents, but only if they can be shown to have created a defective product. Product liability law allows an injured person to sue the manufacturer for design defects, manufacturing defects, or failure to warn. A design defect claim argues that the autonomous system was poorly engineered—for example, if a manufacturer’s lidar system couldn’t reliably detect cyclists, or if the vehicle’s decision-making algorithm prioritized speed over safely stopping for obstacles. A manufacturing defect claim argues that a particular vehicle rolled off the assembly line with a flaw, such as a sensor that wasn’t properly calibrated.

A failure-to-warn claim argues that the manufacturer knew the system had limitations but didn’t adequately communicate them to owners—this is exactly what some plaintiffs alleged against Tesla, arguing that Autopilot was marketed as “self-driving” when it required active human monitoring. One major limitation manufacturers face is that autonomous vehicle technology is genuinely new, and courts may apply different standards for how safe the system should be. If an autonomous vehicle causes fewer accidents than human drivers statistically, some manufacturers argue they should not be held liable for the occasional accident that still occurs. However, the counterargument is that manufacturers shouldn’t market a system as autonomous if it requires human override or monitoring—that’s not truly autonomous. Additionally, if a manufacturer knew about a defect before the accident occurred, failed to issue a recall or software update, and an accident resulted from that known defect, liability becomes much clearer and punitive damages could potentially be awarded.

Owner and Operator Liability for Autonomous Vehicle Crashes

Vehicle owners and operators can be held liable for accidents their autonomous vehicles cause, particularly if the owner failed to maintain the vehicle, ignored warning systems, or allowed unqualified people to operate it. This is where traditional negligence law intersects with autonomous vehicle accidents. If an autonomous vehicle was known to have a malfunctioning sensor, and the owner drove it anyway, the owner—not the manufacturer—might be liable for damages.

Courts have historically held car owners responsible for accidents caused by their vehicles, and that principle hasn’t changed just because the vehicle drives itself some of the time. A major complication arises with vehicles that require human supervision, like Tesla’s Autopilot or other Level 2 and Level 3 systems. If a human operator was supposed to be monitoring the road and keeping their hands on the wheel, but instead was reading a newspaper or sleeping, and the autonomous system failed to avoid an accident, the question becomes: would the accident have been prevented if the human had been paying attention? This is a factual investigation that requires data from the vehicle’s sensors and cameras to determine whether the human’s attention (or lack thereof) was the actual cause. Some courts might find the human operator primarily liable in these scenarios, similar to how a drowsy driver bears responsibility for their crashes.

Insurance Coverage and Liability Limits in Autonomous Vehicle Cases

Insurance is the primary mechanism through which accident victims recover damages, but insurance for autonomous vehicles presents new problems. Traditional auto insurance policies are written assuming a human driver is in control and making decisions; autonomous vehicles violate that assumption. Many insurance companies have begun updating policies to explicitly cover autonomous driving features, but the coverage varies widely. Some policies cap liability at $100,000 to $300,000 in damage per incident, which is insufficient for serious injury or death cases, while others provide higher limits.

Manufacturers like Waymo have purchased dedicated liability insurance for their self-driving taxi fleets, recognizing that the manufacturer—not the vehicle owner—is responsible for the vehicle’s decisions. A critical limitation is that if the accident results from a manufacturing defect, the vehicle owner’s liability insurance might not cover damages because insurance generally doesn’t cover losses from defective products—that’s what product liability insurance is for. This means a victim might have to sue the manufacturer directly through product liability channels rather than filing a claim against the owner’s insurance. Some states have begun requiring autonomous vehicle operators to carry insurance with minimum limits specifically for autonomous driving features, but no national standard exists yet. If a victim is injured by a Waymo self-driving taxi and Waymo’s insurance limits are exhausted, additional recovery might come from suing Waymo directly as an entity, the taxi company that deployed the vehicle, or the software developers involved.

The regulatory landscape for autonomous vehicle liability remains fragmented and incomplete. Federal agencies like the National Highway Traffic Safety Administration (NHTSA) can investigate autonomous vehicle crashes and issue recalls, but NHTSA hasn’t yet established binding liability standards for autonomous vehicles. This means that liability rules are determined largely by state courts applying traditional product liability, negligence, and strict liability principles to new technology. Some states have passed limited autonomous vehicle laws that address testing and deployment, but few have addressed who pays when accidents happen.

A significant warning: the absence of clear legal standards means that liability cases involving autonomous vehicles will likely be litigated vigorously, taking years to resolve and costing enormous amounts in legal fees. Victims and manufacturers alike face uncertainty about how courts will interpret existing law when applied to autonomous systems. For instance, if a autonomous vehicle’s algorithm decides to hit a pedestrian instead of swerving into oncoming traffic (a “trolley problem” scenario), is the manufacturer liable for how the algorithm was programmed to make that choice? Courts haven’t definitively answered this. Additionally, as autonomous vehicles become more common, class action lawsuits might emerge if a particular model or software version is found to have a systemic defect affecting thousands of vehicles—managing liability across such a large group could bankrupt manufacturers or require settlements in the billions.

Comparative Negligence and Shared Fault in Autonomous Accidents

In some autonomous vehicle accidents, fault is shared among multiple parties, and courts apply comparative negligence rules to apportion liability. If a self-driving car failed to detect a pedestrian, but the pedestrian was jaywalking in low light conditions, a court might find both parties partially at fault and reduce the pedestrian’s recovery proportionally. Some states use “pure” comparative negligence, allowing a plaintiff to recover even if they were 99% at fault, while others use “modified” comparative negligence, barring recovery if the plaintiff was more than 50% at fault.

This creates a complication for victims: even if the autonomous vehicle’s manufacturer bears significant responsibility, your recovery might be reduced if the court determines you were partly at fault for the accident. For example, in a collision between an autonomous vehicle and a human-driven car where both drivers failed to yield at an intersection, comparative negligence would divide liability between them. The manufacturer’s insurance and the other driver’s insurance might share the cost, reducing what each source pays.

Establishing Causation Through Vehicle Data and Sensor Records

Proving who caused an autonomous vehicle accident requires detailed technical evidence because it often hinges on what the vehicle’s sensors detected and what the system decided to do. Modern autonomous vehicles record data from cameras, lidar, radar, and other sensors, as well as logs of the vehicle’s decision-making algorithms. This data can show whether the vehicle detected an obstacle, whether it attempted to brake, what speed it was traveling, and what actions it took in the moments before a collision. Unlike human drivers who might have vague memories of an accident, autonomous vehicles create objective records.

However, accessing this data is not guaranteed. Manufacturers often claim that sensor logs and algorithm code are proprietary trade secrets or protected by attorney-client privilege if the data was created during accident investigation. Courts increasingly require manufacturers to disclose this data in litigation, but the process is slow and expensive. If you’re injured in an autonomous vehicle accident, your attorney will need to file discovery motions to force the manufacturer to produce the vehicle’s sensor data and software logs; without this evidence, proving the manufacturer’s liability becomes extremely difficult. In one high-profile incident, investigators had to subpoena Tesla’s data logs to determine whether Autopilot was engaged and whether it attempted to prevent the collision, demonstrating how crucial this technical evidence is.

  • —

Frequently Asked Questions

Can I sue an autonomous vehicle manufacturer directly if their vehicle hits me?

Yes, you can file a product liability lawsuit against the manufacturer if you can demonstrate the vehicle’s design, software, sensors, or safety systems were defective and caused the accident. You’ll need technical evidence showing the autonomous system failed to function as designed.

Is the vehicle owner liable for accidents their autonomous vehicle causes?

Yes, vehicle owners can be held liable for accidents their autonomous vehicles cause, especially if they failed to maintain the vehicle, ignored warning systems, or allowed unqualified people to operate it. However, if the accident resulted from a manufacturer’s defect, liability may shift to the manufacturer instead.

What if the accident was partially my fault?

Depending on your state’s comparative negligence laws, your recovery could be reduced proportionally to your share of fault. In pure comparative negligence states, you can still recover even if you were 99% at fault, though your award would be reduced accordingly.

Will my auto insurance cover an accident with an autonomous vehicle?

Your coverage depends on your specific policy and whether you own the autonomous vehicle or were hit by one. Insurance for autonomous vehicles is evolving; many insurers now offer explicit coverage for autonomous driving features, but policies vary widely in their limits and exclusions.

What evidence proves the autonomous vehicle caused the accident?

Vehicle sensor data (from cameras, lidar, and radar), algorithm logs showing the system’s decisions, maintenance records, and expert analysis of the autonomous system’s performance are key evidence. Manufacturers typically maintain this data, but you may need to subpoena it through legal proceedings.

Can I recover damages if the manufacturer’s insurance runs out?

Yes, you can potentially sue the manufacturer, the vehicle owner, the taxi or deployment company, and any third-party software developers as additional sources of recovery. However, the manufacturer’s liability insurance limits and assets ultimately determine how much can be recovered.

You Might Also Like

Browse more: Accidents | Amputation | Auto | Blog | Brain