Self-Driving Car Accidents: Who Is Responsible When Autonomous Vehicles Crash?

Autonomous vehicles are operating on California roads in growing numbers as manufacturers test and deploy self-driving technology in real-world conditions under permits issued by the California Department of Motor Vehicles. The accidents involving them raise liability questions that existing negligence law was not designed to answer. When a self-driving car causes an injury, fault cannot simply be assigned to the driver as it would be in a conventional collision. The manufacturer of the vehicle, the software developer who created the autonomous driving system, the company operating the vehicle, and even the human occupant may all have potential liability depending on the specific circumstances of the crash and the level of autonomy the vehicle was operating under at the time.

How California Regulates Autonomous Vehicles

California allows the testing and deployment of autonomous vehicles under regulations administered by the California Department of Motor Vehicles. The regulations distinguish between testing with a safety driver present who can take control if needed and fully autonomous deployment without any human driver in the vehicle. Manufacturers must:

• Obtain a testing or deployment permit from the DMV before operating autonomous vehicles on public roads in California
• Report all collisions involving autonomous vehicles to both the DMV and the California Highway Patrol within 10 days of the incident
• Submit annual disengagement reports detailing every instance where the autonomous system failed or encountered a situation it could not handle and a human driver had to take control
• Maintain minimum insurance coverage of 5 million dollars per autonomous vehicle to cover potential injury claims

These collision reports and disengagement reports are publicly available through the DMV and can be valuable evidence in litigation. They provide manufacturer-submitted data about what the autonomous system was doing at the time of the collision, whether there was a human driver present and whether they were monitoring the system, whether any system failures or limitations contributed to the crash, and how the manufacturer categorized the incident. The regulatory framework is evolving rapidly as the technology develops, and any accident involving an autonomous vehicle should be evaluated by an attorney who stays current on both the technology and the applicable regulations.

The Liability Framework: Multiple Potential Defendants

Liability in an autonomous vehicle accident typically runs across 3 categories of potential defendants. The vehicle manufacturer may be liable under product liability theory if a defect in the hardware such as sensors, cameras, or computing systems or a defect in the software that controls vehicle decision-making caused the collision. The software developer, who may or may not be the same entity as the vehicle manufacturer, may be liable independently if the artificial intelligence decision-making system behaved negligently by making choices that a reasonably prudent human driver would not have made. The human operator or owner may be liable if they failed to intervene when intervention was required, if they deployed the vehicle in conditions it was not designed to handle, or if they disabled safety features. Applying the 4 elements of negligence to these facts requires understanding how duty, breach, causation, and damages apply when the primary actor making driving decisions is an algorithm rather than a human being making real-time judgments.

Product Liability vs. Negligence

Product liability claims against vehicle manufacturers do not require proving that the manufacturer was careless in designing or building the vehicle. They require proving that the vehicle was defective and that the defect caused the injury. Defects relevant to autonomous vehicles can be:

• Design defects, where the autonomous driving system as designed cannot handle certain scenarios safely, such as construction zones with temporary lane markings, unusual weather conditions, or pedestrians behaving unpredictably
• Manufacturing defects, where a specific vehicle\'s sensors, cameras, computing hardware, or software malfunctioned due to a production error or installation problem
• Failure to warn defects, where the manufacturer did not adequately inform users about the system\'s limitations, the circumstances where human intervention is required, or the proper way to monitor and override the system

An autonomous driving system that consistently makes incorrect decisions in a specific driving scenario, such as failing to detect pedestrians in low-light conditions, misjudging the speed of oncoming traffic at left turns, or incorrectly categorizing stationary objects as non-hazards, may be defective even if it performs correctly 99.9 percent of the time. The legal standard is not perfection but whether the system is unreasonably dangerous given its intended use and the representations made to consumers about its capabilities.

Negligence claims against manufacturers are also possible when the manufacturer knew of a specific risk through testing or prior incidents and failed to address it before deploying or continuing to deploy the technology. Both product liability and negligence theories can be pursued simultaneously, and the right approach depends on the specific failure mode involved in the accident and what evidence is available about the manufacturer\'s knowledge.

Multi-Party Accidents Involving Autonomous Vehicles

Many autonomous vehicle accidents involve other human-driven vehicles as well, creating liability questions that span both human error and machine decision-making. Our analysis of how liability is determined in multi-vehicle accidents in California provides the legal foundation for understanding how fault is apportioned when more than 2 parties are involved and when some parties are human drivers making real-time decisions while others are autonomous systems operating under programmed parameters.

Autonomous vehicle accident litigation requires technical expertise in addition to legal expertise. You need accident reconstruction specialists with specific experience in autonomous vehicle systems who can interpret sensor data and reverse-engineer what the vehicle detected and how it responded. Software engineers who can analyze vehicle data logs, review the source code if available through discovery, and explain how the AI decision-making system processed the driving scenario are essential expert witnesses. Human factors experts who understand the design decisions behind autonomous driving systems and can testify about whether the human operator had reasonable opportunity to intervene are also frequently necessary in serious cases involving permanent injuries or death.

If you were injured in an accident involving an autonomous or semi-autonomous vehicle, a car accident attorney with experience in technology-related liability cases can identify the right defendants based on the level of autonomy the vehicle was operating under and the specific system failure that occurred, obtain the vehicle data logs and regulatory filings through discovery and public records requests, hire the specialized experts needed to prove your case, and build a claim tailored to the emerging legal standards in this rapidly developing area of law where precedent is still being established.