In Pittsburgh, Pennsylvania, Uber has deployed a small test fleet of self-driving taxis. Anthony, Pittsburgh your Self-Driving Uber is arriving now, Uber (2016), (last visited Jan 22, 2017); Signe Brewster, Uber starts self-driving car pickups in Pittsburgh, TechCrunch (2016), (last visited Jan 22, 2017); Nathan Ingraham, I drove around Pittsburgh in a self-driving Uber, engadget (2016), (last visited Jan 22, 2017). These self-driving vehicles are capable of driving without input from a human driver using artificial intelligence. The artificial intelligence interprets inputs from external LIDAR sensors and reacts by automating responses of the vehicle’s controls. The vehicles transport paying customers as passengers and contain a human Uber driver that oversees the vehicle, taking over the controls if the self-driving features encounter difficulties. Brewster (2016). Several excellent review articles have thoroughly explored the possibilities for future legislation. Comment: “Somebody grab the wheel!”: State autonomous vehicle legislation and the road to a national regime, 97 Marq. L. Rev. 1085. However, now that self-driving vehicles have moved from the realm of science fiction to real roads, the question of liability in a collision becomes a real, near-term concern.
In cases of a collision involving a self-driving vehicle, to whom the liability resulting from the collision is likely to be apportioned is an uncharted area of the law. The causes of action for driver negligence and products liability could potentially be said to apply. Given existing law, it appears likely that both Uber and the human driver could be liable in a self-driving vehicle collision in Pittsburgh.
- Preliminary Policy Guidelines and Absence of Binding Statute
The only current commercial operation of self-driving vehicles is in Pittsburgh, Pennsylvania. Currently, there is no federal statute governing the operation of self-driving vehicles, nor is there a Pennsylvania statute. NCSL (2016), (last visited Jan 22, 2017). Pennsylvania also has no executive order governing self-driving vehicles. Id. The only guidepost available that specifically discusses self-driving vehicles is the Preliminary Statement of Policy Concerning Automated Vehicles from the National Highway Traffic Safety Administration (NHTSA). Nat’l Highway Traffic Safety Admin., U.S. Dep’t of Transp., Preliminary Statement of Policy Concerning Automated Vehicles (2013). This policy breaks down vehicle automation into four levels:
Level 1: Function-specific automation
Level 2: Combined function automation
Level 3: Limited self-driving automation
Level 4: Full self-driving automation
The Uber vehicles on the streets of Pittsburgh are level 3 by this standard. This level is defined as:
Vehicles at this level of automation enable the driver to cede full control of all safety-critical functions under certain traffic or environmental conditions and in those conditions to rely heavily on the vehicle to monitor for changes in those conditions requiring transition back to driver control. The driver is expected to be available for occasional control, but with sufficiently comfortable transition time. The vehicle is designed to ensure safe operation during the automated driving mode. An example would be an automated or self-driving car that can determine when the system is no longer able to support automation, such as from an oncoming construction area, and then signals to the driver to reengage in the driving task, providing the driver with an appropriate amount of transition time to safely regain manual control. The major distinction between level 2 and level 3 is that at level 3, the vehicle is designed so that the driver is not expected to constantly monitor the roadway while driving.
The NHTSA policy does not create any binding rules, but does make recommendations to state drafters of legislation and regulations for level 3 vehicles. Nat’l Highway Traffic Safety Admin., U.S. Dep’t of Transp., Preliminary Statement of Policy Concerning Automated Vehicles (2013).
The NHTSA policy recommends requiring a human supervisory driver employed by a business engaged in testing and using the self-driving vehicle only in that capacity. Id. It also makes several safety recommendations. These include training of supervisory drivers and testing of self-driving vehicle systems before placing them on the road with other drivers. Id. The NHTSA policy also recommends limiting the scope of self-driving vehicle operations to roads compatible with their abilities and minimizing the time required to transition from self-driving mode to driver control. Id. In addition, the policy recommends recording and reporting malfunctions and ensuring that the self-driving systems do not interfere with existing federally required safety features and requirements. Id.
Uber’s self-driving taxi program in Pittsburgh appears voluntarily compliant with these recommendations. Uber provides a “safety driver” that can take manual control of the vehicle at a moment’s notice with a single button press. Ingraham (2016). The self-driving features are additional to the car’s normal features. Id. Additionally, Uber is operating in Pittsburgh with the approval of the mayor of Pittsburgh and staying within the city’s limits. Pittsburgh mayor on driverless cars: ‘This is where the world is going’, WBUR (2016), (last accessed Jan 22, 2017). This voluntary cooperation has earned the project good will, but does not necessarily control liability. Compliance with the NHTSA’s recommendations will likely be argued as evidence that Uber acted with reasonable care in the execution of the program.
However, there is no case law on self-driving vehicle collisions, no statute on point, no state executive order, and no binding federal regulatory policy. Additionally, the artificial intelligence controlling the vehicle introduces a product liability aspect that may complicate the question of duty for the human driver. The first collision in which a self-driving vehicle damages a traditional vehicle, damages property, or strikes a pedestrian will necessarily consider more than a typical case of the same kind.
III. Remedies Available for a Potential Collision Under Existing Common Law
The remedies currently available under tort law that are most likely to apply are negligence and product liability. A suit for negligence could be entertained against the human driver of the taxi or Uber for the driver negligence of its self-driving car systems. A product liability suit could be brought for providing driverless cars, alleging a design defect that renders the cars unable to avoid certain types of accidents.
Pennsylvania applies joint and several liability to any damages in negligence for the owner of a vehicle that knowingly permits an unauthorized driver to operate the vehicle. 75 Pa. Cons. Stat. § 1574 (2016). Ample case law establishes that a negligent, unlicensed, or otherwise unlawful driver controlling the vehicle by permission of the owner creates liability for the owner where the owner had reason to suspect unlawful driving behavior would result. See Burkholder v. Genway Corp., 432 Pa. Super. 36 (Pa. Super. Ct. 1994); Commonwealth v. Tharp, 373 Pa. Super. 285 (Pa. Super. Ct. 1988); Green v. Klein, 2010 Pa. Dist. & Cnty. Dec. LEXIS 444 (Pa. County Ct. 2010); Commonwealth, Dep’t of Public Welfare use of Molek v. Hickey, 136 Pa. Commw. 223 (Pa. Commw. Ct. 1990). The self-driving systems are not a person and thus inherently unlicensed, so their operations would default to the vehicle’s owner. In this case, that would create liability for Uber.
Pennsylvania has comparative negligence by statute and uses the 50% bar rule. 42 Pa. Cons. Stat. § 7102 (2016). This means that negligence on the part of the plaintiff, such as not seizing the wheel and redirecting a driverless car to avoid an accident, does not bar suit unless the negligence is 50% of the fault. Additionally, Pennsylvania does not permit indemnity for common carriers. 74 Pa. Cons. Stat. § 8402 (2016). In other words, there is no way for Uber to cause riders to sign away their right to sue Uber, as part of a waiver, user agreement, or otherwise. Due to these two rules, any cause of action in negligence for damages resulting from a collision involving a self-driving car owned by Uber would involve naming the driver and Uber as defendants and likely primarily recovering from Uber.
Though self-driving cars are new, there are cases concerning a driver’s duty to avoid collisions remaining in effect even when not actively controlling the vehicle. In a case from Pittsburgh in the 1930s, a taxi driver stopped on a trolley track negligently and was then struck by a negligently operated trolley, injuring the passenger. Hughes v. Pittsburgh Transp. Co., 300 Pa. 55, 57-58 (Pa. 1930). The driver ceased controlling the taxi before the trolley approached. The trolley then approached dangerously and made no move to avoid a collision. The cab driver was held to have acted negligently by failing to resume control of the taxi and avoid the oncoming trolley even though the trolley driver could just as easily have prevented the collision. Id. at 58-60. This situation is parallel to how negligence would be assessed for the human driver and the self-driving vehicle systems. The driver’s duty to avoid collisions is not discharged by the self-driving vehicle systems’ ability to do the same. Just as either the cab driver or the trolley driver could have avoided a collision in Hughes, either the human driver or the self-driving car software could avoid a collision. Just as the cab driver was held to have a duty to actively avoid collisions even where the trolley could have avoided the collision for him, the human driver will likely maintain a duty to actively avoid collisions even where the self-driving car could avoid the collision on its own.
In our hypothetical collision, both the self-driving vehicle’s negligence and the driver’s negligence can be applied to Uber so long as the vehicle is acting within the scope of its purpose under the doctrine of respondeat superior. Should a self-driving Uber be used by its human driver or operated by its self-driving features outside of the scope of employment, the question of owner’s liability would then arise. Owner’s liability analysis yields the same answer, as Uber is both the employer and the owner. If negligence analysis is applied, the duty remains with Uber regardless of the approach.
- Products Liability
Damages resulting from a malfunction or systems failure of a self-driving car could alternatively be pursued under the existing common law in products liability. In Pennsylvania, strict liability is applied to a product sold in a defective condition that leads to injury. Gaudio v. Ford Motor Co., 2009 PA Super 102, 531-32 (Pa. Super. Ct. 2009). In Pennsylvania strict liability cases, comparative negligence is not an applicable defense. Gaudio v. Ford Motor Co., 2009 PA Super 102, 540 (Pa. Super. Ct. 2009); Kimco Dev. Corp. v. Michael D’s Carpet Outlets, 536 Pa. 1, 8-9 (Pa. 1993); Dillinger v. Caterpillar, Inc., 959 F.2d 430, 435 (3d Cir. Pa. 1992). However, risk-benefit analysis may be asserted as a defense, in which if the product’s designer can prove that the benefits of the product outweigh the risks the feature is not subject to strict liability. Gaudio v. Ford Motor Co., 2009 PA Super 102, 558-59 (Pa. Super. Ct. 2009).
In the case of the self-driving car, a court might think that an active, human driver at the controls is a necessary element for safety. However, the argument that self-driving cars will cause less collisions overall than cars driven by humans can be put forward. Note: “Look Ma, No Hands!”: Wrinkles and Wrecks in the Age of Autonomous Vehicles, 46 New Eng. L. Rev. 581. This argument, if accepted, may free Uber from product liability exposure in the eyes of the Pennsylvania courts. Gaudio v. Ford Motor Co., 2009 PA Super 102, 558-59 (Pa. Super. Ct. 2009).
Uber is both the employer of the human driver and the owner of the self-driving car, and it is also the manufacturer of the system that makes the cars self-driving. Brewster (2016). If a products liability suit were successfully brought to recover damages after a collision, Uber would still be the party paying for the damages. However, there is a difference between a products liability suit and a negligence suit in such a case.
Products liability suits are more favorable to Uber in two ways. First, by rejecting comparative negligence of the driver the courts reject evidence that the driver could have prevented the accident after the product failure. Dillinger v. Caterpillar, Inc., 959 F.2d 430, 435 (3d Cir. Pa. 1992). Second, the entire case for strict liability can be countered if it is sufficiently shown that the product’s benefits outweigh its risks. Gaudio v. Ford Motor Co., 2009 PA Super 102, 558-59 (Pa. Super. Ct. 2009). This would mean that a collision caused by a design defect would not create actionable strict liability if the self-driving car system is less dangerous than the average human driver. This would remain the case even if the Uber employee behind the wheel negligently failed to take control of the vehicle in time to prevent the collision.
Self-driving cars are now a reality. Tort cases are going to arise from collisions involving self-driving cars soon. Even where there is not an explicit statutory or regulatory binding authority covering such cases, the existing common law tort remedies can be brought to bear. In the example the real world has presented so far, a potential answer is traditional negligence analysis for a common carrier under the doctrine of respondeat superior and/or owner’s liability. An alternative cause of action in product liability is likely to be more favorable to Uber.
When self-driving cars enter private hands, the conflict between the liability of owners, operators, and manufacturers of self-driving cars will be of greater concern. This is especially true where there is no binding authority on point. As the technology continues to advance and proliferate, case law and statutes will likely change to keep pace. In the meantime, existing tort law is not at a total loss even in the face of this radical technological advance. Collisions involving self-driving cars can potentially be handled by traditional products liability and negligence analysis.