Arizona made a deliberate choice to become a testing ground for autonomous vehicles. In 2015, Governor Doug Ducey signed an executive order welcoming self-driving car companies to test on public roads with minimal regulatory barriers. Waymo launched its commercial robotaxi service in the Phoenix area. Cruise operated in Scottsdale. Intel’s Mobileye tested on Arizona highways. The state became, by design, one of the most active AV testing environments in the world.
That decision brought economic investment and technological innovation — and it also put Arizona residents in the path of vehicles operating on software that is still being refined. When those vehicles crash, the question of who is legally responsible does not have a clean answer. And that ambiguity can be devastating for victims who do not know where to turn.
Arizona’s autonomous vehicle framework is codified under A.R.S. § 28-9601 et seq., which defines automated driving systems and establishes that the ADS (automated driving system) is considered the operator of the vehicle when engaged. This is a significant legal statement — it means that when a fully autonomous vehicle is operating under its own system, the traditional concept of a “driver” who can be held personally negligent does not apply in the same way.
What this means practically is that liability shifts from the individual driver to the entity responsible for the ADS — typically the manufacturer, the software developer, or the company deploying the vehicle commercially. Arizona law does not yet have a comprehensive statute that definitively resolves how damages are allocated in AV crashes, which means these cases are litigated under a combination of product liability law, general negligence principles, and agency law.
Product liability. If the autonomous system itself was defective — whether in its design, its manufacturing, or its failure to warn users of known limitations — the manufacturer or software developer can be held strictly liable for resulting injuries. This is the most powerful theory because it does not require proving that anyone was negligent; it only requires showing that the product was unreasonably dangerous and that the defect caused the crash.
Negligence. Even when a vehicle is operating autonomously, the company deploying it has ongoing duties: to adequately test the system before deployment, to monitor its performance, to respond appropriately to known failure modes, and to ensure the vehicle is properly maintained. A company that deploys an AV system it knows has difficulty detecting pedestrians in low-light conditions — and does so anyway without adequate safeguards — may be found negligent regardless of whether the system itself was technically defective.
Human operator negligence. Many AV deployments still involve a human safety driver or remote operator whose job is to intervene when the system encounters a situation it cannot handle. If that person was inattentive, distracted, or failed to intervene when they should have, they and their employer can be held liable under traditional negligence standards. The 2018 Uber AV fatality in Tempe — the first pedestrian death caused by a self-driving vehicle — involved a safety driver who was watching a video on her phone at the time of impact.
In March 2018, an Uber autonomous test vehicle struck and killed Elaine Herzberg as she walked her bicycle across a four-lane road in Tempe. The vehicle’s sensors detected her approximately six seconds before impact but the system failed to correctly classify her as a pedestrian and did not initiate emergency braking. The safety driver was not watching the road.
The case resulted in criminal charges against the safety driver, a civil settlement with the victim’s family, and a temporary halt to Uber’s AV testing program. It also produced a detailed NTSB investigation that revealed systemic failures in how Uber had designed its safety protocols — including a deliberate decision to disable the vehicle’s automatic emergency braking system during testing.
This case established several important precedents for AV liability in Arizona: that the safety driver can face criminal liability for inattention, that the deploying company can face civil liability for systemic design failures, and that the data recorded by the AV system itself is critical evidence in any subsequent litigation.
Autonomous vehicles are among the most thoroughly documented machines ever put on public roads. They continuously record sensor data from LIDAR, radar, cameras, and GPS systems. They log every decision the ADS makes and every input it receives. This data can reconstruct exactly what the vehicle perceived, what it decided to do, and why — in far more detail than any human witness could provide.
The challenge is accessing it. AV companies treat their sensor data and software logs as proprietary trade secrets and will resist disclosure aggressively. Getting this data typically requires litigation, formal discovery, and in some cases, court orders compelling production. This is another reason why having an attorney with experience in AV accident cases — rather than a general personal injury attorney — makes a significant difference in outcomes.
If you are involved in a crash with an autonomous or semi-autonomous vehicle, the steps are similar to any accident — call 911, seek medical attention, document the scene — but with some important additions. Note whether the vehicle appeared to have a human driver or was operating without one. Photograph any visible sensors, cameras, or markings on the vehicle. Note the company name and any identifying numbers on the vehicle.
Contact an Arizona self-driving car accident attorney immediately. The data preservation window in AV cases is just as narrow as in truck accident cases, and the companies involved have legal teams that will begin managing the narrative immediately. Call Phillips Law Group at (602) 222-2222 for a free consultation.
Available 24/7 
