The True Cost of Bringing Fully Automated Vehicles to Market: An Analysis of the Liability Implications to Come

in Government/Manufacturing/Public Policy/Technology/Volume V

By Sharifa Hurt

The introduction of fully autonomous vehicles (AVs) to the auto market is expected to yield many benefits. Crash causation studies have found that human drivers were a primary cause of the over 5 million car crashes in the U.S. annually.[1] Autonomous vehicles may greatly reduce driver recognition error, the factor that is the root cause of human driver crashes, ultimately resulting in safer roadways.[2] Fully autonomous vehicles are also expected to provide many additional benefits, such as increasing the mobility of seniors, reducing traffic, gas emissions, and costs associated with accidents, and increasing individual productivity/free time.[3] All these anticipated benefits bring forth exciting news for tech investors, consumers and the general public at large. But regulatory uncertainty poses a set of very complex and unresolved liability questions for the autonomous vehicle industry. Who assumes liability when a fully AV rideshare car collides with a car operated by a human driver? What about a collision between two fully AV? Or if an AV collides with a jaywalking pedestrian?

The Society of Automotive Engineers (SAE) is a global association of more than 128,000 engineers and experts in the aerospace, automotive and commercial vehicle industries, As a leader in the automative engineering industry, it has identified six levels of car autonomy.[4] The scale ranges from levels 0-5, referring to a range of functionality from complete driver control to complete self-driving autonomy.[5] [6] At Level 1, “Driver Assistance”, the car is equipped with technologies designed to assist the driver, such as cruise control. At Level 2[7], “Partial Driving Automation”, there is a function of at least two simultaneous driver assistance systems in place. For example, steering and acceleration/deceleration programs are on, with the expectation that the human driver is responsible for all remaining driver tasks.[8] At Level 3, “Conditional Driving Automation”, the system handles all aspects of driving while the driver is expected to be available to take over driving if requested. [9] At Level 4, “High Driving Automation”, the system is fully responsible for driving tasks within limited-service areas and does not require the intervention of a human driver.[10] At Level 5, “Full Driving Automation”, the vehicle is capable of performing all driving functions under any conditions. [11] No cars with technology above a “Level 2” autonomy function[12] are currently available for consumer purchase in the United States today. [13]

Autonomous vehicles currently in the consumer market in the United States are incapable of handling the full range of driving conditions and require a human driver. Accordingly, most states still hold the driver responsible for following traffic laws and liable for any resulting accidents. The liability implications for fully automated vehicles are still largely in flux. However, a thorough analysis of liability theories suggests tech companies and auto manufacturers should prepare to take on a daunting and unpredictable increase in liability with fully automated vehicles.

The Emergence of the AV Market

The total global investment in autonomous vehicle (AV) technology is currently in excess of $200 billion, and this figure is expected to grow exponentially in the immediate future.[14] Big players in the tech and automotive industry have begun testing AV technology in real road conditions.[15] The current focus on the initial phase of driverless vehicles is on the mobility-as-a-service (MaaS), which consists of fleets that would be used for public services such as rideshare and delivery.[16] This is an exciting development for consumers. The increase in transportation accessibility will positively impact many large populations that rely on public transport, such as elderly, disabled, and low-income communities. However, the resulting anticipated shift of driving from private to public also impacts accident liability.

A resident near Google’s driverless car division, Waymo, recently reported troubling observations stemming from their real road testing of fully AV technology.[17] She reported observing “what looked like a glitch in the self-driving software: the car seemed to be using her property to execute a three-point turn…sometimes a few of the SUVs would show up at the same time and form a little line, like an army of zombie driver’s-ed students.”[18] Fortunately, that particular system glitch did not result in any collisions. But, what if one of the vehicles backed into the resident’s unsuspecting pet? In an accident with a human driver, it is obvious that the driver would be held liable. But who is held liable when the driver is a machine?


One tort theory that will likely be pursued in the event of auto accidents involving a fully AV and a human driver is negligence. According to the Legal Information Institute at Cornell University, if a plaintiff can prove that a defendant failed to behave with the level of care that someone of ordinary prudence would have, they have a claim for negligence.[19] The following four elements are required for a plaintiff to prevail on a negligence theory of liability: (1) the existence of a legal duty that the defendant owed to the plaintiff; (2) defendant’s breach of that duty; (3) plaintiff’s sufferance of an injury; and (4) proof defendant’s breach caused the injury.[20]

In the case of an accident involving a fully AV and a human driver, the plaintiff would have to prove that the AV owner owed the other human driver a duty of care, that they breached that duty of care, and that the breach of the duty is what resulted in the accident. Today, in similar cases that involve auto accidents between two human drivers, the issue of whether either driver breached a duty of care that resulted in the auto accident is very heavily contested. In jury trials, liability often largely depends on which driver’s account of the accident the jury finds more trustworthy. Each party has an equal opportunity to thoroughly explain their perspective of the accident to the jury in the case of an auto accident between two human drivers. However, in the case of an accident between a fully AV and a human driver, a fully AV can’t testify to aspects of the accident that may suggest liability on behalf of the human driver. Its’ video recordings may offer some evidence as to what exactly transpired, but that is assuming that the video recordings capture every angle of the crash, which does not seem realistic. Thus, the perspective of the human driver will likely be much more persuasive.

Additionally, machines commonly malfunction. Tragedies resulting from machine malfunctions have been highly publicized. There have recently been several highly publicized malfunctions which resulted in large tragedy and devastation. For example, there was a system glitch in the Boeing 737 Max and 737-800 aircrafts which resulted in hundreds of fatalities. [21] In the aftermath of tragic machine malfunctions and accounts of people witnessing fully AV technology malfunction in real road testing as described earlier, it is relatively easy for people to accept the fault of a machine malfunction being the cause of an accident generally. 

A jury will likely have a much more difficult time finding that extremely technologically advanced robots acted reasonably in the circumstances surrounding an accident than human drivers do. In addition, the fact that a robot cannot testify to their account of the accident presents further challenges for the AV industry’s chances or prevailing at trial. Thus, even without a right to a strict liability claim, a human plaintiff’s chances at prevailing on a general negligence theory will likely be very high.

Product Liability

While a plaintiff’s chances of prevailing on a general negligence claim can be high under the above-mentioned theory, the most likely theory of liability plaintiffs will pursue against auto manufacturers will be product liability. According to the Legal Information Institute at Cornell University, the legal theory of product liability refers to “the liability of any or all parties along the chain of manufacture of any product for damage caused by that product.”[22] This theory of liability encompasses the manufacturer of component parts, the assembling manufacturer, the wholesaler, and the retail store owner.[23] For a plaintiff to prevail against an auto manufacturer on a products liability theory, the plaintiff would have to prove that: (a) the defendant sells a product the plaintiff uses; (b) the defendant is the commercial seller of the product; (c) the plaintiff suffered an injury; (d) when the defendant sold the item, it was defective; and (e) the defect was an actual and proximate cause of the plaintiff’s injury.[24]

In the case of an auto accident involving a human driver and a fully AV, the plaintiff would have to explicitly prove some defect in the automobile caused the accident. This theory of liability creates the potential to hold many more parties liable for the accident. While this description of proof sounds very similar to proving negligence (and negligence is a theory on which product liability can be proven), product liability is much more sweeping than general negligence. For example, through pursuing a product liability claim, one can make claims that go beyond how the automation was negligent at the time of the accident, or about how the product’s overall design is flawed and is therefore at fault for the accident.  If a plaintiff can establish there is a hypothetical alternative design that would be safer than the original design, which is as economically feasible and practical as the original design, they will prevail on a design defect theory of liability. The plaintiff must also demonstrate that the design retains the primary purpose behind the original design despite the changes made.

In the context of fully AVs, the design defect theory imposes a lot of liability on a variety of entities because there are an infinite number of design improvements one could think up that may prevent an auto accident in any given circumstance. For example, if a pedestrian jaywalks across the street and is hit by a fully AV, it is easy to imagine a design defect theory of liability being pursued. The plaintiff will try to prove that the automation design is flawed because if the automation had a feature that brought the car to a halt sooner, the accident could have been prevented. Again, the plaintiff will have the benefit of an emotional testimony and their perspective of the accident will likely be the dominant narrative, so it is easy to foresee the plaintiff’s side being very persuasive. Contrast this scenario with a case in which a human driver was operating the vehicle in the same circumstance, and the added burden for the AV defendant is more clearly demonstrated. The plaintiff would likely have a more difficult time proving that the human defendant operated the car, abided by traffic and safety laws, and was at fault, as opposed to the alternative version of events consisting of the human plaintiff running out into the street and exercising very reckless judgment.

The AV industry faces unique challenges in prevailing on product liability claims against auto manufacturers. These challenges are further exacerbated by the current state of regulatory uncertainty on fully AV technology. It is likely that juries will be particularly skeptical about a manufacturer’s failure to include the technology that may have prevented an accident. There are an infinite number of features that a plaintiff could consider may prevent an accident in any given circumstance. In addition, the lack of eyewitness testimony on the part of the AV industry defendant, the heavy advertising of how technologically advanced/reliable AVs are by tech companies, and general skepticism regarding fully AV’s being newly introduced to market, make the AV manufacturer’s defense increasingly challenging.

Third Party Claims

In addition to the many theories of liability that fully AVs are prone to impose upon the AV industry, there will also likely be a vast number of third-party entities exposed to liability. One of the benefits of fully AVs is the potential for providing more public transit and infrastructure opportunities. This is because vehicle automation is expected to increase the use of automated rideshare and shuttle fleets.[25] This shift in public transit and infrastructure will potentially change the need for individualized parking spaces and lots, which could dramatically transform land use, yielding various environmental net benefits and reducing traffic.[26] Automated driving systems also have the potential to increase mobility for seniors and people with disabilities, and expand transportation options for underrepresented communities.[27]  Fully AVs are expected to be transformative in many industries. However, this transformation brings forth looming questions of third-party liability.

While there has been scarce regulation on fully autonomous vehicle technology, there haven’t been any regulatory safeguards proposed to protect third parties from liability. So what happens when a self-driving rideshare vehicle collides with a curb, or a pedestrian, or even another fully autonomous rideshare vehicle? Without additional regulatory protections, an affected party can hold several various parties liable, as they can pursue any of the liability theories previously mentioned. If the rideshare vehicle was owned by Uber, for example, there are theories on which the plaintiff could potentially hold Uber, the auto manufacturer, the retailer where the car was purchased, and given that any of the public transit models for fully AVs have been adopted, even the city liable. Today, when similar types of collisions occur with a human driver, most claims don’t reach beyond the driver and the rideshare company. However, with fully AV technology operating in a human driver’s place, the liability that is removed from the human driver will likely be redistributed amongst third party entities. 

Pre-Allocation of Liability for AVs

In the attempt to mitigate the increased burden of liability resulting from the industry shift from private to public drivers, tech companies and AV manufacturers may reduce the risk of liability through pre-allocation. Manufacturers, suppliers, and sellers may be able to create some predictability in risk and/or liability by using disclaimers, warranties, and indemnification provisions. A product is less likely to be subject to product liability claims if it is covered by warranties and disclaimers because in California, and most jurisdictions, the manufacturer has the ability to limit warranties.[28] In effect, a manufacturer can avoid/limit manufacturing defect claims in contracting with the selling entity. However, particularly in the early stages of introducing fully AV technology to market, testing has indicated the big issue is likely to be with algorithm glitches, as was previously described. Because glitches with the technology are likely to fall under design defect liability, mitigating risk through warranty and disclaimer will not protect AV industry defendants from what is likely to be the largest source of liability risk, which is design defect liability.

Recent Cases Involving “Level 2” Autonomy Function

While there isn’t currently any case law regarding fully AVs, due to their lack of availability on the United States market, there has been litigation regarding crashes involving “Level 2” autonomy function. The National Highway Traffic Safety Administration (NHTSA) reports at least 18 fatalities resulting from “Level 2” driver autonomy technology.[29]  In 2017, there was an accident in northern Florida when a Tesla model car, operating on Autopilot, crashed into a semi-truck that the car’s sensors failed to detect. [30] The National Transportation Safety Board (NTSB) found that the car failed to send a visual or audible warning to the driver to put his hands back on the steering wheel. [31] Accordingly, the NTSB recommended that Tesla limit where the autopilot feature can be activated in order to ensure that drivers pay more attention while using the feature. [32] The NTSB subsequently criticized Tesla for failing to adopt this recommendation in 2021.[33]

In Palm Beach County Florida, there is currently a similar trial underway,  in which a Tesla Model 3 collided with the underbelly of a tractor-trailer shortly after the driver.[34] The crash was fatal, so there will likely be no eyewitness testimony, but an investigation by the NTSB found that the driver assistance feature did not detect the truck crossing the highway.[35] This trial, set for February of 2023, will be the first of numerous Autopilot collision trials, and its verdict is expected to say much about the future of liability for fully AVs.[36] Whether the Autopilot feature creates a false impression of fully AV capabilities in the mind of the driver may have significant ramifications in the way of product liability for the fully AV industry. The extent to which the jury is prone to be more sympathetic to a human driver in opposition to AV technology will also be on display.


The introduction of fully AVs to market is expected to be a revolutionary auto industry advancement. This development in the automotive space is anticipated to yield numerous net benefits to the consumer. Fully AV technology is expected to solve numerous decade long issues in the safety, environmental, and mobility spaces. However, in bringing resolution to these spaces in the form of fully AV technology, there is a shift in automotive function from private to public driver. 

This shift is economically beneficial to the consumer on its face. The shift of auto accident liability from the consumer to the tech industry giant seems like a major economic victory. However, increased liability costs bring with them increased responsibilities for public safety on behalf of the AV industry. Thus, in the context of tech giants and automakers stepping into the realm of public safety, it is important to bear in mind the reality that corporate business models function based on cost-efficiency. Consideration of the recent firsthand accounts of AV testing, the current lack of regulatory oversight, and transportation entities’ failure to appropriately deal with tech malfunctions, indicates that this industry breakthrough will likely have a revolutionary impact in the legal space and ominous implications for public safety.

[1]Traffic Safety Facts, Critical Reasons for Crashes Investigated in the National Motor Vehicle Crash Causation Survey (February 2015).

[2] “Automated Vehicles for Safety.” NHTSA. Accessed October 4, 2022.


[4] “J3016_202104: Taxonomy and Definitions for Terms Related to Driving Automation Systems for on-Road Motor Vehicles.” SAE International. Accessed October 4, 2022.


[6] Id.

[7] Id.

[8] Id.

[9] Id.

[10] Id.

[11] Id.

[12] Id.

[13] Supra note 2.

[14] Yeruva, Vinod. “Council Post: Autonomous Vehicles and Their Impact on the Economy.” Forbes. Forbes Magazine, February 16, 2022.

[15] Id.

[16] Id.

[17] Chafkin, Max. “Self-Driving Cars Are Starting to Look Like a $100 Billion Bust.” Bloomberg law. Accessed October 7, 2022.

[18] Id.

[19] “Negligence.” Legal Information Institute. Accessed October 4, 2022.

[20] Id.

[21] Chengevelyn. “Boeing 737 Passenger Jet Crashes in China with 132 People on Board.” CNBC. CNBC, March 22, 2022.

[22] “Products Liability.” Legal Information Institute. Accessed October 6, 2022.

[23] Id.

[24] Id.

[25] Supra note 2.

[26] Id.

[27] Id.

[28] Cal. Com. Code § 2313 (Deering, Lexis Advance through Chapter 425 of the 2022 Regular Session).

[29] Nayak, Malathi. “Tesla Fatal-Crash Lawsuit to Test Musk’s Autopilot Claims.” Bloomberg Law, 13 Sept. 2022.

[30] Id.

[31] Id.

[32] Id.

[33] Id.

[34] Id.

[35] Id.

[36] Id.