LIABILITY FOR CAR ACCIDENTS IN A (MOSTLY) DRIVERLESS WORLD

Author: Benjamin Fooks, Senior Editor   

  Imagine you are diving your car, and your GPS suggests you drive home a different way than your typical route.  Being a creature of habit, you ignore your GPS and follow your normal route. Now imagine a similar scenario, but in this situation, you are in an autonomous car which begins to take you home a different way than normal.  You instinctually turn off the self-driving feature and continue to drive on your own.  Now imagine you get hit going through an intersection, an accident which otherwise would not have happened while using the self-driving capability of your car.  How is liability divided when you intentionally turned off your cars’ self-driving features? 

A.    Different levels of Self-Driving Cars

     Potential liability is impacted by the level of autonomy of the car in question.  The Society of Automotive Engineers have created a six-level ranking of autonomy for self-driving cars.[1]  At level 0 the car requires a human driver in order to function, and the human is in control of all primary functions at all times.[2] Level 1 is best depicted as a car with adaptive cruise control.[3]  Level 2 vehicles can control the speed or steering of the car, but the car still mainly relies on input from a human driver.[4]  Level 3 vehicles are cars that can perform most vital functions on their own, but they rely on human drivers to stay alert to conditions on the road and step in as needed.[5]  Level 4 automated cars are able to perform all major functions on their own while also monitoring road conditions.[6]  Some level 4 vehicles may even function without humans under the right conditions.[7]  Level 5 cars transcend the need for human intervention, converting humans into simple cargo.[8]

B.    Suggested Approaches to Liability in Self-Driving Car Cases

     Some legal scholars have already suggested blanket approaches to apply to the liability of autonomous cars.  Most existing liability proposals question how liability should be divvied up when autonomous features are engaged.  Some of these schemes might still apply where the driver has chosen to disengage their cars’ autonomous features. 

i.                 Manufacture Default /Strict Liability

     Some scholars have suggested taking a strict liability approach when self-driving cars are involved in accidents.[9]  This would place liability on manufacturers of self-driving cars as opposed to the car passenger-supervisor.[10]  This approach to liability only works in a scenario where the autonomous car has failed to work as the manufacturer intended.[11]  This would not directly address an accident where the self-driving capabilities had been disabled.  Thus, if the opposing plaintiff attempted to sue on a theory of strict liability they could be barred from recovery because they are using the car in a fashion that the manufacturer did not intend.[12]

ii.               Autopilot in Airplanes

     Another potential approach to liability is to treat autonomous cars like the autopilot in airplanes.[13]  When the aircraft autopilot system causes a crash it is treated as a products liability case with the autopilot manufacturer liable for any injury.[14]  This approach to strict liability could potentially translate to autonomous cars.[15]  A majority of aircraft crashes are due to human error, particularly on landing and takeoff when autopilot systems are disengaged.[16]  This avoids any potential strict liability since the pilot is liable for the negligent operation of the aircraft.[17]  For example, in Klein v. United States, Klein neglected to turn on autopilot while landing a Boeing 727 aircraft.[18]  The Court held Klein to be the sole and proximate cause of the crash.[19]  Klein provides great insight into the scenario at issue since the driver therein turned off his self-driving features.  Under this precedent, then, if the passenger/supervisor were the defendant in a case where they turned off the self-driving features on their car, they would be liable for negligence.    

iii.             Safety Features

     Another possible approach is to treat the autonomous features as safety features.  Autonomous cars have the potential to drastically increase the safety of the drivers and passengers inside the vehicle.[20]  If autonomous driving is a safety feature, then how safety features have been dealt with in the past could predict how courts will rule on this scenario.  There are several reasons to view autonomous features as safety features: their ability to reduce the number of drunk drivers from the road, control the speed on roadways, and generally reduce the likelihood that accidents will occur on the road. 

     First, self-driving features should be viewed as safety features for their ability to give potential drunk drivers a way of getting home without endangering other motorists on the road.[21]  The National Highway Traffic Safety Administration’s (NHTSA) 2007 study found that twenty eight percent of all accidents on the road are caused by drunk driving.[22]  If autonomous cars reduce the average percentage of drunk driving crashes then it should be considered a safety feature.    

     Next, self-driving features could act as a safety feature by controlling a vehicle’s speed.  The 2007 NHTSA study found that speeding accounts for around thirty percent of all roadway accidents.[23]  Eliminating thirty percent of accidents alone would be a reason to consider self-driving features to be safety features.  

     If self-driving cars could remove the impetus for over fifty percent of all accidents on the road by reducing speeding and drunk driving then autonomous driving systems would be more than a safety feature, it would be a safety revolution.[24]      

     In addition to decreasing the likelihood of drunk drivers being on the road, and controlling the speed of a car, computers react faster than human brains, this could potentially lessen the yearly number of accidents.[25]

C.    Future Liability for Drivers of Non-Autonomous Vehicles

     With a possible safety revolution on the horizon, it is likely that people electing to drive their cars without the input of a self-driving system could face harsher penalties.

i.                 Comparative Negligence

     If a driver were hit by someone after disabling the autonomous features on their vehicle, they might run into some comparative negligence issues which could lower their overall recovery.  

a.      Seatbelt Defense

One example of how recovery could be diminished is the seatbelt defense, a defense which lessens recovery in tort for not following safety precautions.[26]  To use this defense, a defendant, must prove that a plaintiff’s failure to wear a seatbelt contributed to their injuries.  The failure of a plaintiff to wear a seatbelt is then considered when determining the plaintiff’s comparative negligence.[27]  Seatbelts are a safety feature in vehicles which decrease a passenger’s or driver’s odds of being seriously injured or killed.[28]  This is such an important safety feature that most states have developed laws mandating the use of seatbelts.[29] 

The results of Tesla’s autopilot system have shown to reduce the likelihood of accidents by 40 percent.[30]  With such a drastic decrease in accident likelihood, the development of a defense akin to the seatbelt defense would make sense for the scenario at issue.  If the seatbelt defense stands as a model for the self-driving car defense, then potential plaintiffs who deactivate their autonomous driving features could have their recovery discounted by the damages they would not have sustained had they been using the self-driving feature on their cars.[31]

b.     Other Comparative Negligence Models

     The seatbelt defense is not an available defense in all fifty states, for example, in Ohio, a defendant cannot use the failure to wear a seatbelt to prove comparative negligence.[32]  However, not wearing a seatbelt is examined to reduce the compensatory damages award for non-economic losses.[33]  Some states will likely develop their own framework for dealing with comparative negligence when drivers disengage their autopilot features.[34] 

ii.               Calculating Fault Driving Non-Autonomous Car

One potential hurdle to this proposed autonomous car defense is the difficulty in calculating the fault of the driver for disengaging the self-driving features on their car.  For example, a driver may have turned off the autonomous features on their car to avoid a collision with another driver.  In this situation, it might be difficult to argue that the driver breached a duty to themselves when they turned off their car’s autonomous features to avoid a perceived threat of imminent harm.   

iii.             DUI Accident Insurance Model

An example of how driving without use of self-driving features may impact insurance coverage can be illustrated by drunk driving instances.  The insurance provider Progressive will cover an accident even if the accident was the result of a DUI.[35]  However, DUI drivers are considered riskier by insurance agencies, and as a result they must pay more for coverage.[36]  In some instances, insurance companies may refuse to insure a person if they have had enough DUIs or if their history is so severe that they are uninsurable.  At Progressive, insurance rates increase by 13 percent after the first DUI.[37] 

The danger of electing not to use the self-driving features on your car could be comparable to the danger of driving intoxicated.  Drivers with a blood alcohol content above .08 are three times more likely to get in an accident than sober drivers.[38]  Some reports suggest that the Tesla Autopilot could reduce the likelihood of accidents by 40 percent.[39]  If this statistic holds true then it is possible that auto insurers will treat driving without the use of autopilot systems similar to the way they treat DUI drivers.  This could lead to higher premiums and possibly being classified as uninsurable for electing not to use self-driving features. 

D.    Conclusion

     Autonomous cars, no longer relegated to the world of science fiction, are becoming more and more a reality.  In the future, most cars on the road could be autonomous.  As a result, there is a need to develop a more robust tort law for the different potential claims which will arise in the world without drivers.

     Autonomous cars are not infallible and will ultimately be involved in accidents.  Autonomous driving features should be viewed as safety features on vehicles, and refusal to use autonomous features should be handled similarly as other safety features.  If a driver is injured in an accident more than he would have otherwise been injured had he used the autonomous features on his car, his recovery should be reduced.  Eventually, drivers of nonautonomous cars will also likely see higher insurance rates than people with self-driving cars since they will be considered riskier to insure.[40]  

     As a nation we have a vested interest in protecting the safety of our citizens.  The advent of autonomous cars provides an avenue to drastically increase the overall safety on our roadways.

1 Rachel Theodorou, "With Cars Like These, Who Needs Policies?" - the Inevitable Battle Between Autonomous Vehicles, the Insurance Industry, Manufacturers and Consumers, 35 Syracuse J. Sci. & Tech. L. 72, 75 (2019). (Theodorou analyzes each level of autonomous vehicles in greater detail to show that the ultimate end to the development of self-driving cars could be the end of car insurance providers).  

2 Id at 76.

3 Id.. at 76 

4 Id.. at 76  

5 Id at 77..

6 Id at 78..

7 Rachel Theodorou, "with Cars Like These, Who Needs Policies?" - the Inevitable Battle Between Autonomous Vehicles, the Insurance Industry, Manufacturers and Consumers, 35 Syracuse J. Sci. & Tech. L. 72, 7875 (2019).

8 Id at 79..

9Jeffrey K. Gurney, Sue My Car Not Me: Products Liability and Accidents Involving Autonomous Vehicles, 2013 U. Ill. J.L. Tech. & Pol'y 247, 2573 (2013) (Gurney endorses a products liability approach to self-driving cars when they are involved in crashes.  Gurney hypothesizes that since the autopilot systems controlled the car that they were most likely at fault for the accident, and by an extension, the manufacturer would also share in the blame.) 

10 Id.

11 Id.Jeffrey K. Gurney, Sue My Car Not Me: Products Liability and Accidents Involving Autonomous Vehicles, 2013 U. Ill. J.L. Tech. & Pol'y 247, 253 (2013).

12 Williams v. Brown Mfg. Co., 45 Ill. 2d 418, 425, 261 N.E.2d 305, 309 (1970) citing (Restatement (Second) of Torts § 402A (1965)).

13 Kyle Colonna, Autonomous Cars and Tort Liability, 4 Case W. Reserve J.L. Tech. & Internet 81, 93 (2012). (Colonna looks at many things similar to self-driving features in cars in order to predict how tort liability will be decided.  She looks at not only the autopilot in airplanes, but also the autopilot on submarines and how liability is divided when they are involved in accidents.

14 Id. at 97.

15 Id.

16 Id.

17 74 A.L.R.2d 628 (Originally published in 1960) (this ALR entry compiles cases where the human element on an aircraft caused an accident and where the pilots were found to be either negligent or not.  This demonstrates that in airplane crashes pilots can be held negligent even if they heavily rely on autopilot.).

18 Peter Y. Kim, Where We're Going, We Don't Need Drivers: Autonomous Vehicles and Ai-Chaperone Liability, 69 Cath. U. L. Rev. 341, 348 (2020) citing (Klein v. United States, 1975 U.S. Dist. LEXIS 15348, at *13 (D. Md. Nov. 11, 1975)).

19 Id.

20 Gary Marchant, Rida Bazzi, Autonomous Vehicles and Liability: What Will Juries Do?, 26 B.U. J. Sci. & Tech. L. 67, 85 (2020) (Marchant briefly discusses the safety features innate in self-driving cars in his article trying to figure out how juries will decide on the topic.).

21 Frank Douma & Sarah Aue Palodichuk, Criminal Liability Issues Created by Autonomous Vehicles, 52 Santa Clara L. Rev. 1157, 1163 (2012) (Douma analyses the possible benefits of drunk individuals utilizing autonomous vehicles to get from place to place as opposed to drinking and driving).

22 NHTSA’S NAT’L CENTER FOR STAT. AND ANALYSIS, Traffic Safety Facts (2007) (This report shows data about accidents and their causes from 2007).

23 Id. NHTSA’S NAT’L CENTER FOR STAT. AND ANALYSIS, Traffic Safety Facts (2007)

24 Id.

25 Supra note 13Kyle Colonna, Autonomous Cars and Tort Liability,, at 4 Case W. Reserve J.L. Tech. & Internet 81, 93. (2012).

26 65 Am. Jur. Proof of Facts 3d 1 (2002) (This American Jurisprudence entry explains the seatbelt defense as it applies to comparative negligence in tort liability)

27 Id.

28 Jesse N. Bomer, The Seatbelt Defense: A Doctrine Based in Common Sense, 38 Tulsa L. Rev. 405, 406 (2002) (Bomer analyses the seatbelt defense and its current status in the United States.  They go overHe reviews the effectiveness of seatbelts in histheir analysis). 

29 Id.

30 Hasan Siddiqui, Gone in Sixty Seconds: Fading Automobile Insurance Costs in A Driverless Future, U. Ill. J.L. Tech. & Pol'y, Spring 2018, at 221, 236.

31 65 Am. Jur. Proof of Facts 3d §1 (2002).

32 Ohio Rev. Code Ann. § 4513.263 (West 2022).

33 Id.

34 Id.

35 Auto Ins. After a DUI, PROGRESSIVE, https://www.progressive.com/answers/dui-and-insurance/ (last visited Aug. 15, 2021) (This page details the repercussions for people convicted of driving while under the influence of alcohol). 

36 Id.

37 Id.

38 Supra note 30.Hasan Siddiqui, Gone in Sixty Seconds: Fading Automobile Insurance Costs in A Driverless Future, U. Ill. J.L. Tech. & Pol'y, Spring 2018, at 221, 236

39 Christopher Ingraham, How Just a Couple Drinks Make Your Odds of a Car Crash Skyrocket, W.A. Post (Feb. 9, 2015) https://www.washingtonpost.com/news/wonk/wp/2015/02/09/how-just-a-couple-drinks-make-your-odds-of-a-car-crash-skyrocket/  

40 Auto Ins. After a DUI, PROGRESSIVE, https://www.progressive.com/answers/dui-and-insurance/ (last visited Aug. 15, 2021)Supra note 35.