Journal of Law, Information and Science
|
|
Home
| Databases
| WorldLII
| Search
| Feedback
Journal of Law, Information and Science |
|
Unmanned Vehicles and US Product Liability Law
COMMENT BY STEPHEN S WU[*]
Unmanned vehicles may someday become commonplace in transportation, in industry, and for a wide variety of applications. Just as automobiles and trucks displaced the horse, unmanned vehicles may become the mainstay for transporting cargo, and autonomous vehicles may displace human-driven automobiles and trucks. When unmanned and autonomous vehicles become commonplace, the legal system will need to address liability arising from the accidents that will occur with the use of these vehicles. With its sophisticated system of product liability statutes, doctrines, and guidance, the law of unmanned vehicles in the US may have worldwide influence on the scope and nature of unmanned vehicle liability.
Commentators may produce interesting predictions of future legal developments regarding unmanned vehicles. Nonetheless, a look at past case law regarding automatic devices and robots also provides useful insights on the future of unmanned vehicle liability. Courts may face a growing number of novel cases as a result of rapid and accelerating technological developments. They may need to weigh the risk of danger from particular vehicle designs against the benefits of such designs before legislatures can consider possible legislation or regulatory oversight to maintain the safety of unmanned vehicles. Moreover, the trajectory of case law to date provides useful trends to help predict future legal developments. Finally, reviewing past decisions yields intriguing possibilities of future arguments that may someday break new ground in litigation regarding unmanned vehicles.
I have already published two earlier versions of this paper. First, I created an initial version to support the discussion of robotics liability during the 2010 American Bar Association Annual Meeting program entitled ‘When Good Robots Do Bad Things: Responsibility and Liability in an Era of Personal & Service Robotics.’ The American Bar Association Section of Science & Technology Law’s Artificial Intelligence and Robotics Committee presented the program on 6 August 2010 in San Francisco. Next, I revised the paper for a program entitled ‘Sudden Acceleration into the Future: Liability from Autonomous Driving and Robotics’ presented at the 51st Annual Meeting of the Association of Defense Counsel of Northern California and Nevada on 9 December 2010. This version of the paper incorporates additional cases and analysis regarding unmanned vehicles.
The premise of the paper is to create a summary and overview of the kinds of cases seen to date concerning the liability arising from the use of unmanned vehicles and robots generally. It is not intended as an exhaustive listing of all cases involving the actual or possible liability of the manufacturers of unmanned vehicles or robots. Instead, it provides a sample of typical categories of cases against manufacturers and owners of autonomous devices. I have also included other cases that raise interesting points about liability for failing to use robots. In addition, I discuss cases against parties other than manufacturers as a means of exploring the liability of manufacturers if they had been sued. I conducted the research for this paper on a legal research service, using search terms related to various theories of liability, variations of the word ‘robot’, and variations of the word ‘autopilot.’
Reviewing autopilot cases is useful, because vehicles with autopilots use a form of automation. Autopilot automation is a less advanced form of the kind of automation we see and will continue to see in unmanned vehicles, especially autonomous vehicles. Courts may decide that precedents regarding vehicles with autopilots apply by analogy to unmanned vehicles.
Part I highlights robotics and autopilot cases concerning issues of liability under traditional theories of products liability, such as strict liability and negligence. Part II discusses cases shedding light on defences used in robotics and autopilot cases. Part III discusses some cases deciding procedural matters in suits relating to robotics and autopilot liability. The facts of these cases pose interesting scenarios for possible liability despite the fact that the courts rendering these decisions never reached the substance of the disputes. Part IV serves as a discussion of liability generally, noting patterns in the cases and drawing conclusions.
Unmanned vehicle and robot manufacturers may face product liability lawsuits raising a number of theories, including strict liability, negligent design, negligent failure to warn, and breach of warranty. Research did not reveal a large number of decided cases involving liability issues under these theories. Nonetheless, the cases revealed by the search raise product liability issues typical of machinery used in industrial manufacturing settings. Moreover, some cases raise product liability issues involving autopilot systems on planes and boats.
A plaintiff asserting a strict liability claim against an unmanned vehicle or robot manufacturer must plead and prove, under a typical state’s law, that the defendant sold a product that was defective and unreasonably dangerous at the time it left the defendant’s hands, the product reached the plaintiff without substantial change, and the defect was the proximate cause of the plaintiff’s injuries. Under a negligent design theory, a plaintiff would seek to show a robotic or unmanned vehicle manufacturer had a duty to exercise reasonable care in manufacturing the machine, the manufacturer failed to exercise reasonable care in making the machine, and the defendant’s conduct proximately caused the plaintiff’s damages.
Jones v W + M Automation, Inc,[1] is in many ways a typical product liability case. A piece of equipment struck the plaintiff when he entered into an area behind a safety fence, and the main issue in the case concerned the existence of a genuine issue of material fact, on summary judgment, as to whether the system was defective when the defendant sold it. The equipment at issue, however, was a robotic gantry loading system used in a General Motors auto plant.
GM had purchased the system and then installed it without an interlock system intended to stop the machine when people are present, thereby allowing employees to work on the system within the danger zone behind the safety fence while the system was operating. Gripper arms of the system hit the plaintiff in the head while he was standing behind the safety fence and he became pinned against a pedestal, injuring his head. The Occupational Safety and Health Administration fined GM for not installing an interlock system and for allowing employees to work behind the safety fence while the system was operating. The plaintiff sued the manufacturers, among other things, under theories of strict liability, negligence, failure to warn, and breach of warranty.
The court held that summary judgment was appropriate for the component manufacturer defendants under the ‘component part’ doctrine, which states that a manufacturer of a non-defective component part of a product is not liable if its part is incorporated into another product that might be defective. The defendants were also entitled to summary judgment because the plaintiff failed to introduce evidence in opposition to summary judgment showing that the system was defective. GM’s modifications to the system were the apparent cause of the accident. The plaintiff’s expert said that the system failed to meet robotics standards of the American National Standards Institute (ANSI), but the court found that the system comported with voluntary industry standards, and the ANSI standards did not apply to the system at issue.
Similarly, other cases have found no liability when plaintiffs failed to adduce evidence of a defect. Payne v ABB Flexible Automation, Inc,[2] (decedent crushed by a robot’s gripper arm in an auto wheel plant while working within the ‘cell’ in which the robot operated, with some evidence of user error). In Payne, the court said that the failure to meet ANSI standards, and the problems with unexpected movements, admitted by the defendant manufacturer, were irrelevant, because these problems did not cause the accident. Accordingly, it affirmed summary judgment in favour of the manufacturer.
Likewise, in an airplane autopilot case, Ferguson v Bombardier Services Corp,[3] the plaintiffs could not succeed in their claims against an autopilot manufacturer, because they had no evidence at trial to prove defects in the autopilot. The plaintiffs brought suit following a catastrophic crash of an Army National Guard plane against the manufacturers of the plane and the manufacturers of the autopilot. Before trial, the district court excluded the testimony of the plaintiffs’ two experts concerning defects in the autopilot, and thus, during trial, the plaintiffs had no evidence of a defect in the autopilot to support claims against the autopilot manufacturer. Consequently, the district court entered judgment for the autopilot manufacturer as a matter of law, and a jury later found for the remaining defendants.[4] The appellate court affirmed the exclusion of the testimony and the judgment in favour of the defendants.
The case turned on the cause of the accident. The plaintiffs pointed to various defects in the autopilot system. The defendants claimed that the aircraft was improperly loaded, and that when a pilot left the cockpit to walk to the rear of the aircraft, the aircraft became unstable, and a gust of wind caused the plane to lose control.[5] The Court of Appeals held that it was proper to exclude an expert the plaintiffs called to testify as to the defects in the autopilot, because on cross-examination at a pre-trial hearing, the expert admitted that the information from the plane’s flight data recorder was equally consistent with the defendants’ theory of improper loading as with the plaintiffs’ theory of autopilot defects, and thus his testimony would not assist the jury in understanding the evidence.[6] The plaintiffs’ other expert was to testify about Federal Aviation Regulations that supposedly required the manufacturer to include a warning annunciator with the autopilot system, but the expert could not show that the regulations actually contained such a requirement.[7]
In another airplane autopilot case, a federal appeals court affirmed a judgment for defendants in a negligence and strict liability case against the manufacturer of a small airplane, the airplane’s seller, and the engine manufacturer. In Moe v Avions Marcel Dassault-Breguet Aviation,[8] the plaintiffs pointed to a defect in the autopilot as a cause of the crash, among other causes, saying that if the pilots failed to disengage the autopilot, no warning would sound when the pilots attempted to control the plane manually. In fact, the plaintiffs claimed the autopilot would cause the plane to act against the pilots’ manual control in this situation. The plaintiffs pointed to this factor as one of the causes of the accident.[9]
The jury, however, found none of the defendants liable where the plaintiffs could not prove a cause of the accident involving the small airplane. The Court of Appeals held that the jury’s answers to special interrogatories on the verdict form were internally consistent, since neither party convinced the jury of the cause of the accident, and thus the plaintiffs failed to establish negligence or causation by a preponderance of the evidence.[10] The Court of Appeals also held that the trial court’s instructions were proper.[11] Accordingly, the Court of Appeals affirmed the verdicts for the defendants.
Similarly, in Glorvigen v Cirrus Design Corp,[12] a fatal airplane crash case, the court found that the plaintiffs (trustees for the deceased pilot and passenger) could not support a strict liability claim against an airplane manufacturer, because they presented no evidence of a manufacturing defect and could not proceed under strict liability on the basis of inadequate instructions. The plaintiffs’ main claim was that Cirrus voluntarily took on the duty to train the decedent pilot on the use of the autopilot system and failed to deliver the proper training to the pilot, including a failure to provide certain follow-up training. Their theory was that the crash occurred because the pilot did not know how to use the autopilot system when weather required flying by instruments. Thus, the plaintiffs did not allege a manufacturing defect. Since the court did not permit the plaintiffs to proceed in strict liability under an inadequate instructions theory, the court entered summary judgment for Cirrus on the strict liability claim.[13]
Moreover, the plaintiffs’ implied warranty claim was barred by a disclaimer of implied warranties in the purchase agreement for the plane.[14] In addition, the plaintiffs could not support an express warranty theory based on the failure to provide follow-up avionics training.[15] The court therefore granted summary judgment on both the implied warranty claim and the express warranty claim, holding that negligence was the only means of proceeding for the plaintiffs.[16] As discussed in Section 1.2 below, however, the Court of Appeals of Minnesota later reversed a verdict for the plaintiffs under the negligence theory, holding that Cirrus owed the pilot no duty to provide follow up training.[17]
Boucvalt v Sea-Trac Offshore Services, Inc,[18] involved claims of gross negligence against an autopilot manufacturer, whose system allegedly caused a yacht accident. The owner and passengers of the yacht brought suit after it struck a well jacket owned by Chevron. The suit included claims against Raymarine, which manufactured the yacht’s autopilot system.[19] Raymarine moved for summary judgment to seek dismissal of the plaintiffs’ claim for punitive damages under general maritime law.
In opposition, the plaintiffs pointed to deposition testimony of witnesses who claimed the autopilot’s fluxgate compass acted erratically in the presence of large metal objects, Raymarine refused to test the compass separately from the rest of the autopilot system, and Raymarine knew about the problems with the compass.[20] The Court of Appeal, however, held that Raymarine’s alleged conduct did not rise to the level of ‘reckless or callous disregard for the rights of others, or gross negligence’ and therefore did not support a punitive damages claim.[21] The decision did not reach the merits of the negligence claim against Raymarine, but instead simply cut off punitive damages as one possible remedy.
Provenzano v Pearlman, Apat & Futterman, LLP,[22] provides an interesting ‘case within a case’ in that it concerned a malpractice claim against a law firm that represented the plaintiff in a previous case against the manufacturer of a robotic television camera. The camera had hit her in the head in a television studio, where she worked as a hair stylist. Following a defence verdict in the underlying case, the plaintiff sued her attorneys for malpractice. The court held that the plaintiff’s expert report did not raise a genuine issue of material fact to show that she would have prevailed in the underlying suit and granted summary judgment. The report failed to explain why the accident stemmed from a design defect, as opposed to recent repairs on the camera by the studio.
Three other cases pose interesting issues regarding the standard of care for robotics manufacturers as compared to the standard of care for human operators of machines. For instance, Arnold v Reuther,[23] involved a driver who hit a pedestrian while making a left turn in his car. The court, affirming dismissal of the suit, held that the defendant driver did not have the ‘last clear chance’ to avoid the accident, because he could not have prevented the accident after the plaintiff darted out onto the street. In so holding, the court stated:
A human being, no matter how efficient, is not a mechanical robot and does not possess the ability of a radar machine to discover danger before it becomes manifest. Some allowance, however slight, must be made for human frailties and for reaction, and if any allowance whatever is made for the fact that a human being must require a fraction of a second for reaction and then cannot respond with the mechanical speed and accuracy such as is found in modern mechanical devices, it must be realized that there was nothing that Reuther, a human being, could have done to have avoided the unfortunate result which the negligence of Mrs. Arnold brought upon herself.[24]
This decision raises the possibility that once we have autonomous vehicles, the courts will raise the standard of care for manufacturers to avoid collisions, since robots can act faster and more accurately than humans. Moreover, if humans’ driving record is worse than typical autonomous vehicles someday, it may even be negligent for humans to drive themselves. The more general question is whether someday failing to use a robot may be negligence.
Indeed, in one case, a court found sufficient evidence of a defect in a conventional ventilator used to supply oxygen to patients, because of testimony that the device used outdated technology and could have included a redundant backup system and robotic monitoring system. That is, the lack of such systems made the machine unreasonably dangerous.[25]
Likewise, in Mracek v Bryn Mawr Hosp,[26] a patient sued a hospital following manual surgery that resulted in alleged damages, which the human doctors performed after the hospital’s ‘da Vinci robot’ malfunctioned and could not perform the surgery. Although it is not clear in the reported decision, the plaintiff apparently alleged that the hospital’s doctors negligently performed the surgery. The plaintiff in that case, however, also sued the manufacturer for providing a malfunctioning device that could not perform the surgery with the precision he needed. In other words, the plaintiff apparently believed that a functioning robot would have performed the surgery better than the human doctors. The decision concerned the manufacturer’s liability after voluntary dismissal of the hospital.
Although the failure to use a robot may someday create liability, the flip side is that humans may still have a duty of care to avoid accidents, even after delegating some of the operation to a robot. This liability may arise in the use of semi-autonomous machines, like autopilots. In Brouse v United States,[27] the court held that a pilot had a duty to be on the lookout to prevent air-to-air collisions while flying under ‘robot control’ (evidently, autopilot). The Army ‘Black Widow’ fighter struck the plaintiff’s Aeronca Cub after the pilot failed to notice it.
Likewise, one admiralty case, Shaun Fisheries, Inc v Lanegan,[28] found no defect in an autopilot, but apportioned liability to the operator for, among other things, failure to keep a lookout. In Shaun Fisheries, the fishing boat F/V Shaun had the right of way and proceeded straight, while the tug Mary Catherine and her tow, the barge Bandon, also proceeded straight, and the boats headed toward an eventual crossing and collision. Mary Catherine’s owner claimed that the autopilot on the Shaun malfunctioned, causing a sudden swerve, although the suit did not include the autopilot manufacturer.
The court found no defect in the autopilot, especially in light of expert testimony that the autopilot was engaged and the position of the rudder was amidships and only slightly canted to port, suggesting that the Shaun was on autopilot, going more or less straight, just as would be expected for a vehicle on a straight course with a functioning autopilot.[29] Despite the lack of any defect in the autopilot, the court found each vessel 50% at fault, because the Shaun’s captain did not keep a proper lookout, or avoid an accident once it became apparent that the Mary Catherine would not yield.[30] One possible cause was that the Shaun was on autopilot and the operators failed to remain alert enough to disengage the autopilot once the craft were apparently on a collision course. As a result, the owner of the Mary Catherine was able to shift half the liability to Shaun’s owner and captain. In other words, the claim against the operators for failing to disengage the autopilot was more successful than its attempt to blame the autopilot. Moreover, the human lookouts bore some of the liability for the crash.
An unmanned vehicle or robot manufacturer may be held liable for a negligent failure to warn if the manufacturer knows or has reason to know that the product is likely to be dangerous for its intended use, it has no reason to believe that users will realise its dangerous condition, and it fails to exercise reasonable care to inform users of its dangerous condition. Nonetheless, if the danger is open and obvious, the manufacturer has no duty to warn users.
For instance, Jones,[31] discussed above in Section 1.1, addressed the plaintiff’s failure to warn claim. The court held that the danger involved with the plaintiff of going behind the safety fence while the system was in operation was an open and obvious one. Accordingly, the court held that summary judgment should have been granted to defendants on the failure to warn theory.
Also, the Glorvigen[32] plane autopilot case discussed in the previous section, involved a failure to warn claim after the federal court’s summary judgment decision. The federal court later remanded the case to a Minnesota state court for lack of jurisdiction. The state trial court then held a jury trial resulting in a verdict in favour of plaintiffs. The plaintiffs’ theory was that Cirrus had a duty to provide adequate instructions concerning the autopilot system on the plane, and Cirrus’ failure to provide sufficient autopilot instructions caused the crash.
The Court of Appeals of Minnesota reversed the verdict in favour of the plaintiffs.[33] The court held that Cirrus had no duty to provide follow-up training to the pilot regarding the autopilot system. Accordingly, the plaintiffs’ negligent failure to warn theory failed.[34] The court also held that the plaintiffs’ negligence claims against Cirrus sounded in educational malpractice, and that such claims are barred as a matter of law.[35] The court cited public policy reasons to bar educational malpractice claims, including the lack of a standard of care, uncertainties about causation because students’ conduct may be an intervening cause of damages, the possible flood of litigation against educational institutions, and the desire to avoid the need for court oversight over the daily operations of educational institutions.[36]
Research revealed cases involving claims against the operators of vehicles using autopilot equipment. A passenger seeking damages against an airline successfully obtained reversal of a trial court judgment against her in Nelson v American Airlines, Inc.[37] The plaintiff, Mary Nelson, was a passenger on an American flight to Los Angeles. Shortly after the plane reached its cruising altitude, the pilots turned on the autopilot to hold the plane’s altitude, but the autopilot caused the plane to suddenly dive. Nelson claimed damages from injuries allegedly sustained during the sudden dive.[38]
In the absence of any objection, the trial court applied the doctrine of res ipsa loquitur, which raised an inference of negligence by the airline. The Court of Appeal held that since American was a public carrier, it could rebut the inference of negligence by proving that it exercised the utmost care and diligence through proof that it did not cause the accident or that an unknown and unpreventable cause resulted in the accident.[39] The Court of Appeal said that a defect in the autopilot could have caused the accident.[40] Nonetheless, the Court of Appeal found no evidence to eliminate negligent maintenance as a cause of the action and therefore reversed the trial court’s judgment in American’s favour.[41] Interestingly, Nelson had not sued the manufacturer of the autopilot or the airplane, but nonetheless, obtained reversal of the judgment against her.
Similarly, in the Shaun Fisheries[42] barge collision case discussed in the previous section, the barge owner had mixed results in blaming a fishing boat’s autopilot for a collision. As discussed above, the court found no defect in the autopilot but found liability against the operator of the fishing boat, apparently for failing to keep a lookout and to disengage the autopilot and avoid the collision.
One of the common types of case I found on my research concerned the limitations placed on the ability of an injured employee to sue his or her employer due to the exclusivity of recovery under the workers’ compensation system. Employees cannot generally sue their employers for workplace injuries if state law says that workers’ compensation is the exclusive means of recovering compensation for such injuries. Nonetheless, an exception exists if the employer’s conduct rises to the level of intentional conduct, in which case an employee can seek compensation in excess of workers’ compensation by suing his or her employer.
I found three workers’ compensation cases involving the use of robots: Miller v Rubbermaid Inc;[43] State ex rel Scott Fetzer Co v Industrial Comm’n of Ohio;[44] and Edens v Loris Bellini, Spa.[45] In Miller, the court affirmed the trial court’s entry of summary judgment in favour of the employer where the record showed that the decedent employee placed himself in danger of being crushed by a robot used in plastic injection moulding, the employer did not know of any dangerous condition of the machine, there was no evidence of safety violations, and the employer had not inadequately trained the decedent. Under these circumstances, there was no genuine issue of fact regarding intentional conduct by the employer.
In Scott Fetzer, the Supreme Court of Ohio affirmed an order denying a writ of mandamus to an employer after the Industrial Commission of Ohio approved additional workers’ compensation for an employee injured by a robotic die cast machine that closed and severely injured the worker’s upper body. The employee’s job was to remove bad parts stuck in the die from the danger zone of the machine. The employer had removed safety controls on the machine and when the die unexpectedly closed, the worker sustained injury. The robotic device that normally removed good parts from the die did not always work, requiring workers to manually remove bad parts from the danger zone. The court affirmed a finding of safety violations in light of these circumstances.
In Edens, the plaintiff’s decedent was struck and killed by a robotic shuttle used to transport wool to and from dye vats. The decedent’s co-workers had disconnected safety mats that stopped the shuttle if someone stepped on them. They disconnected the mats because the repeated stopping of the shuttle ‘aggravated’ the shuttle operator. When the decedent was checking for leakage from the vats, a co-worker activated the shuttle, which hit and killed him. The court affirmed a dismissal of the employer and co-worker defendants for lack of subject matter jurisdiction, in light of the court’s finding that the decedent was a ‘statutory employee’ subject to workers’ compensation laws, the exclusivity of workers’ compensation, and the lack of evidence that the employer or co-workers intended to harm the decedent.
A plaintiff suing an unmanned vehicle or robot manufacturer must prove that the defect or failure to warn was the proximate cause of the plaintiff’s damages. Sometimes, the intervening act of a third party breaks the causal chain between the defendant’s act or omission on one hand, and the plaintiff’s damages on the other. For instance, misuse or modification[46] of the product may constitute an intervening cause.
Three of the cases discussed above involved causation issues. In the da Vinci medical robot malfunction case, the court found no evidence in the record that the robot’s malfunction and failure to operate caused the injuries resulting from surgery conducted by human doctors.[47] Payne, which involved a robot crushing a worker in its ‘cell,’ held that plaintiff failed to provide evidence that a programming error or lack of a safety feature was a proximate cause of the worker’s injuries.[48] In the television camera case, Provenzano, the court found that the plaintiff failed to prove that defects within the television camera caused it to strike the plaintiff, as opposed to negligent repairs.[49]
The question of causation may arise if a plaintiff tries to sue a manufacturer where, but for an automated machine, the plaintiff would not have been injured, but it is difficult to say that the machine was a substantial factor in the injury. In the following cases, the plaintiff did not sue the manufacturer, but if it had, questions of causation would have arisen.
For instance, in Leister v Schwans Sales Enters, Inc,[50] the plaintiff pickup truck driver was injured while bringing fencing materials to the defendant’s pizza facility for the purpose of building a fence around a palletizing robot. An employee of the defendant hit the pickup with his delivery truck. The court held that the plaintiff was a statutory employee of defendant and his claims were barred because of the exclusivity of workers’ compensation.[51]
Imagine, though, if the plaintiff had sued the manufacturer of the robot. He could claim that, but for the robot, he would not have been injured. He was in his pickup truck to build a fence around the robot. Had the defendant not purchased the robot, the plaintiff would not have been on site to build the fence, and he therefore would not have been injured. In that sense, the robot ‘caused’ the accident. Nonetheless, such a claim would almost certainly fail, because the robot did not meaningfully contribute to the accident. Moreover, the conduct of the defendant’s employee driving the truck was a superseding cause of plaintiff’s damages.
Likewise, in Romano v Browne,[52] the court reversed an order denying a landlord summary judgment for premises liability after a woman tripped on a power cord used to recharge a mail robot in the mail room. The court blamed the tenant for leaving the cord exposed and found no defect in the premises. If the plaintiff had sued the manufacturer of the robot for having a power cord that could trip people, he or she would also find it difficult to prove a design defect caused the injury. Machines require power cords, and the danger of tripping is obvious. The owner of the device, the tenant in this case, has control of the device and where to place it while recharging so as to avoid hazards to those walking by. Thus, the tenant’s conduct would be a superseding cause of the injuries.[53]
The robotics and autopilot cases I found raise five defences to liability claims that will affect liability from the manufacture and use of unmanned vehicles and robots. These defences, of course, are only a subset of the possible defences to a products liability claim. The most common defence was the workers’ compensation exclusivity defence, which bars an employee’s claim against an employer for workplace injuries, including those sustained during the use of a robot. In the workers’ compensation cases discussed above, the plaintiff was unable to overcome this defence.[54] Nonetheless, in Scott Fetzer, the employer could not avoid enhanced compensation for the employee in light of the employer’s safety violations.
Likewise, Behurst v Crown Cork & Seal USA, Inc,[55] did not foreclose an intentional tort claim against an employer. The case arose out of an accident involving a blank transfer robot (moving metal from die to die) used in a can-making plant. The plaintiff’s decedent was killed when, apparently, she was trapped in the danger zone when the machine restarted and the access door closed behind her. The court found a jury question as to an intentional tort claim against an employer in light of:
• The employer’s knowledge of the flawed performance of the machinery and a history of prior accidents.
• The employer’s alleged refusal to reprogram the machine.
• The employer’s tolerance or encouragement of unsafe maintenance practices.
• The employer’s insistence on understaffing its production line.
• The employer’s placement of the decedent alone in the production line without sufficient training.[56]
The second defence raised by these cases is the component parts doctrine. Under this doctrine, the manufacturer of a non-defective component part is not liable if it is incorporated into a defective product that causes injury to the plaintiff. In Davis v Komatsu America Industries Corp,[57] the Tennessee Supreme Court answered a certified question from the US Court of Appeals for the Sixth Circuit and held that Tennessee law recognises the component parts doctrine. The case involved a Sharp plant that manufactures microwaves. The system stamps out metal parts, and robots transfer the parts from press to press. Following a stoppage in the line, the plaintiff was removing a piece of metal waste when a co-worker restarted the line and a press injured the plaintiff’s hand. The manufacturer of the equipment was not liable since the machinery left its hands in a non-defective condition, and the employer had disabled a safety sensor in order to allow the equipment to operate. The alleged defects in the equipment did not concern the robots, but rather the stamping equipment.
In Jones[58] discussed above, some of the defendants were component manufacturers that provided non-defective products. The court affirmed summary judgment in favour of these manufacturers.
Jones also involved a third defence — subsequent modification of the equipment. The employer installed the robotic system without an interlock safety system. Such a system could have stopped the machine while the plaintiff was in harm’s way. Accordingly, the manufacturers were not liable.
The fourth defence appearing in the cases is the government contractor defence. The US Supreme Court recognised the government contractor defence in Boyle v United Technologies Corp.[59] The defence immunises suppliers of equipment to the government from state law liability for design defects in the equipment where the federal government approved reasonably precise specifications for the equipment, the equipment conformed to those specifications, and the supplier warned the federal government about the dangers in the use of the equipment known to the supplier, but not the government.[60] The purpose for the defence is to protect the exercise of the federal government’s discretion in designing equipment it needs and to prevent circumvention of the limitations on federal government liability in the Federal Tort Claims Act.[61]
A federal district court applied Boyle to immunise an aircraft manufacturer and autopilot manufacturer, following a catastrophic crash of an Air Force plane.[62] The plane crashed after a sudden loss of control and loss of electrical power. The plaintiffs’ theory was that the autopilot malfunctioned, causing a loss of control, and other defects prevented the pilots from recovering. The plaintiffs sued the manufacturer of the aircraft and of the autopilot (among other defendants) under theories of strict liability, negligence, and breach of warranty.[63]
The court held that the airplane manufacturer and autopilot manufacturer were immune from suit under the government contractor defence. The Air Force, in an exercise of its discretion, created specifications for the autopilot and other parts at issue, there was no dispute that the parts in question conformed to Air Force specifications, and the Air Force was warned about safety issues arising from the specifications.[64] Accordingly, the case was ‘precisely the kind of case to which the government contractor defence was intended to apply.’[65] Manufacturers of unmanned vehicles and robots may be able to take advantage of the government contractor defence when they are building them to government specifications and warn the government of design deficiencies.
The final defence appears in Housand v Bra-Con Industries, Inc.[66] In that case, a mechanical arm on an assembly line for GM minivans struck the plaintiff while he was cleaning an oil spill around the machines during a work break. Someone restarted the line, and the accident occurred. The court granted summary judgment because of, among other things, the ‘sophisticated user’ defence. Under this defence, the manufacturer is not liable for supplying a product to a knowledgeable user who has reason to know of any dangerous condition in the product. The court held that GM was a sophisticated user, GM was closely overseeing the use of the machines, and therefore the manufacturer had no duty to prevent or remedy any alleged defect.[67]
Some unmanned vehicles and robotics cases involve claims of product liability, but the court decisions dealt largely with procedural issues. These cases are helpful in showing the factual scenarios that may give rise to liability for manufacturers. Some of these cases, though, do not shed much light on substantive questions of liability.
The claims against the manufacturer in Behurst[68] discussed above, turned on a procedural issue. Even though the court denied summary judgment to the employer for the plaintiff’s intentional tort claim, the court granted summary judgment to the manufacturer. The court granted the manufacturer’s motion, because of a ten-year statute of repose that requires a product liability claim to be brought within 10 years after the first purchase, and the machine in question had been purchased 12 years before the suit was filed.
Adams v Gates Learjet Corp,[69] presented another case turning on a time limitation. The plaintiffs were owners of Learjet Model 24 planes who sued for the cost of retrofitting equipment on the Model 24 planes, which the Federal Aviation Administration (FAA) found to be deficient and in violation of applicable regulations. The FAA issued an airworthiness directive, and the plaintiffs incurred costs, including downtime, to fix the deficiencies. Moreover, some owners sold their aircraft at substantial losses, because they could not afford the cost of the modifications. The plaintiffs sought only economic damages since they found out about the deficiencies before any crash occurred.
The court held that, assuming plaintiffs had tort claims, their claims were barred by the two-year Texas tort statute of limitations. The last plane was delivered on 27 April 1979. On 13 June 1983, the FAA issued a proposed airworthiness directive on the Model 24’s deficiencies, and the final directive was published in the Federal Register on September 6, 1984, more than two years before the plaintiffs brought suit on 4 October 1986. Thus, even if the statute did not start to run until the plaintiffs had constructive notice of the defect in September 1984, their October 1986 suit was barred by the two-year statute of limitations.[70] Moreover, the court held that any warranty claim would have a limitation of four years, starting on the date of sale. Since the last delivery was in April 1979, any warranty claim would be barred as of April 1983, and the October 1986 complaint was thus time-barred.[71]
This case is interesting because plaintiffs sought only economic losses associated with fixing the deficiencies in the autopilot system of the airplane. Such suits are possible for defects in unmanned vehicles in the future, before any accident occurs, but the economic loss doctrine may limit plaintiffs to proceeding under contract theories, such as breach of warranty. The Adams court did not reach the merit of plaintiffs’ claims, but regulatory directives and findings of violations may help plaintiffs proceed under a negligence per se theory (if a tort claim is viable) or under laws against unfair and deceptive trade practices.
In Bynum v ESAB Group, Inc,[72] the Michigan Supreme Court reinstituted a defence verdict in a case involving an injury to a plaintiff that was operating a robotic welding system. The court did not detail the nature of plaintiff’s claims against the defendant manufacturer. Rather, the case involved the plaintiff’s claims of alleged racial biases of some of the jurors, which the court rejected.
Some cases involved questions of jurisdiction. For instance, Bou-Matic, LLC v Ollimac Dairy, Inc,[73] involved defective robotic dairy milking machines that allegedly did not work properly and failed to work as represented. Unfortunately, the court does not describe the nature of the alleged defects.
In any case, the California dairy company filed a state court complaint against the Wisconsin manufacturer and the California dealer. The Wisconsin manufacturer then filed a diversity action in federal court against the California dairy and California distributor seeking a declaration of rights. Diversity jurisdiction in US federal courts exists where the parties are citizens of different states in the US or the proceedings involve a citizen’s case against a non-resident alien. The question in Bou-Matic was whether the parties should be realigned to make the dairy the plaintiff and the manufacturer and dealer defendants, which would destroy diversity jurisdiction. In light of the facts and circumstances, the court did, in fact, realign the parties and dismiss the case for lack of jurisdiction.
Rodriguez v Brooks Pari-Automation, Inc,[74] concerned a robot the plaintiff was installing in an elevator shaft. While the plaintiff was in the shaft to communicate with co-workers, another co-worker activated the robot, which descended rapidly towards plaintiff. The plaintiff moved his body out of the way of the robot, but it severed his thumb while it passed him.
The question in Rodriguez was whether the Texas resident plaintiff fraudulently joined the Texas building owner, Texas Instruments (TI), in order to defeat diversity jurisdiction. After TI removed the case to federal court, the plaintiff moved to remand. The court held that the plaintiff might have viable claims against TI for premises liability and under a contractual duty to control a contractor and make sure the contractor does the work in a safe manner. The court found that plaintiff had not fraudulently joined TI, granted the motion to remand, and did not reach the plaintiff’s strict liability and negligence claims.
Aynesworth v Beech Aircraft Corp,[75] is another removal case involving strict liability and negligence claims against a plane manufacturer following a crash of a Beech airplane. Beech brought third party claims against the companies involved with the autopilot system in the plane under the theory that a malfunction in the autopilot system caused the pilot to become disoriented and lose control of the plane. The plaintiff’s theory, however, was that the plane crashed because of a malfunction in the right propeller, which would have placed liability entirely on Beech.[76]
The plaintiffs originally filed suit in a Texas state court and, after a trial that resulted in a jury deadlock and mistrial, Beech removed the action to the US District Court for the Western District of Texas on the basis of diversity of citizenship. The case turned on whether the Texas defendants defeated diversity, in light of the Texas residency of the plaintiffs. The court held that the Texas defendants were not nominal parties, and the plaintiffs did not abandon their claims against the Texas autopilot companies, even though in closing argument, plaintiffs’ counsel urged the jury to reject Beech’s contention that an autopilot malfunction caused the accident. The plaintiffs’ counsel said, as an alternative to Beech’s liability under the propeller theory, that if the jury were to find the autopilot companies liable, they expected a verdict in their favour.[77] The Court also held that Beech waived its right to remove, because it went all the way through trial in state court and then, only after a mistrial failed to produce the result Beech wanted, removed the case to federal court.[78]
Accordingly, the Texas defendants defeated diversity jurisdiction and the court remanded the case to the Texas state court. Since the case turned on jurisdiction, the court did not deal with the merits of the claims against the autopilot companies. Consequently, the only result on liability reported by the court was that the state court jury could not decide whether or not liability existed, whether on the propeller theory or the autopilot theory.
Finally, some reported cases concerned the preclusive effect of prior judgments without reaching the merits of the dispute. For instance, in Casey v Palmer Johnson Inc,[79] the plaintiff boat owner sued a company that repaired his boat, claiming that the defendant failed to repair a defective autopilot. The plaintiff alleged that the day after he picked up his boat, he put it on autopilot, and its defect caused the boat to run aground.[80] In a previous collection action, Palmer Johnson sued Casey for the unpaid bill on the repairs. In that action, Casey claimed that the repair work was inadequate.[81] The court held that the parties had not previously litigated the effectiveness of the repairs on the autopilot, and therefore Casey was not barred by res judicata or collateral estoppel from claiming defective repairs to the autopilot in a later action.[82] The court never reached the merits of Casey’s claim of defective autopilot repairs.
Likewise, in United States v Athlone Industries, Inc,[83] the court reviewed a summary judgment in favour of a defendant manufacturer of robotic pitching machines in a case brought by the federal government following an earlier suit brought by the US Consumer Product Safety Commission (CPSC). The court described the defects as follows:
In our semi-robot age, as a substitute for the batting practice pitcher, inanimate machines have been manufactured which confront the player in the batter's box. In this case, some of the machines were defective and more wild than an erratic pitcher. In fact some of the machines were mysterious and unpredictable; even when disconnected from their power source, these machines retained such a high degree of tension in the spring and cable that with the slightest vibration, the pitching arm would unexpectedly swing forward and downward at great speed, striking any unsuspecting person within its range, allegedly causing injuries that were as serious as fractured skulls and loss of eye sight.[84]
Unfortunately, the court did not describe the alleged defects in any more detail and did not decide the merits of the government’s claims of defects. Instead, the case turned on whether the prior CPSC action barred the government’s action seeking civil money penalties under res judicata. The court held that the district court’s summary judgment for the defendant based on res judicata was in error, because the two suits involved different conduct, different wrongs, and different evidence.
As a side note, the court stated in passing that the government brought suit against the manufacturer ‘[s]ince robots cannot be sued.’[85] The court’s assumption that robots cannot be sued is certainly true today. Nonetheless, this assumption may not hold forever, as robots become more sophisticated, intelligent, and autonomous.
The cases identified in the research show a number of noteworthy patterns. First, a substantial number of the reported cases involve robots in auto plants and those used in die press or moulding equipment used to stamp out parts. It makes sense that a large number of reported cases involve these kinds of robots, because both auto/auto parts manufacturers and companies creating metal or plastic parts using die press/moulding equipment have been early adopters of robotics, the equipment used in these industries is highly dangerous to employees, and accidents involving this kind of equipment can be catastrophic. I expect that a substantial portion of robotics litigation will continue to arise from these industries in the short run. Nonetheless, over the long run, as personal and service robots enter the mainstream and are on sale in the mass market, the share of cases comprised of industrial accident in these industries will likely decline relative to cases involving robots purchased by consumers. Many more cases will arise from malfunctions or defects in personal and service robots, as the consuming public begins to use them in large numbers.
Second, I found a substantial number of cases involving alleged defects in autopilot devices. Again, it makes sense for parties to litigate these cases following catastrophic crashes and the loss of life, given the stakes involved. If unmanned vehicle cases are similar to autopilot cases, we will likely see products liability litigation arising from catastrophic crashes. The autopilot decisions to date, however, seem to focus on classic issues in products liability cases, such as what was the real cause of the accident, and whether the device had any defects. I expect the courts to analyse unmanned vehicle cases in the same way.
Third, a substantial number of cases in this research involved efforts by employees to obtain compensation from employers beyond that offered by workers’ compensation. State workers’ compensation systems provide compensation without the need for litigation against the employer, but also limit compensation. It is not surprising, therefore, that employees injured in accidents involving industrial robots seek to obtain extra compensation. Again, however, the percentage of robotics cases involving workers’ compensation will decline as personal and service robots become mass-market products and consumers start bringing cases against manufacturers.
Finally, these cases reflect our current times, in which unmanned vehicles and personal and service robots are not commonplace household or consumer items. The cases raise some interesting possibilities that standards of care will change and that the failure to use a robot for certain tasks may give rise to liability. In the Mracek da Vinci surgical robot case, for instance, the bad surgical outcome allegedly occurred because the robot was not used to complete the surgery. As robots become better at precision or dangerous work than humans, we are likely to see more cases in which plaintiffs blame humans for not using robots. Likewise, as autonomous vehicles become better drivers than humans, we are likely to see cases in which plaintiffs blame humans for not using autopilot functions or fully autonomous vehicles.
Ultimately, the future is open as to the types of unmanned vehicle and robotics liability claims we will see in future decades. As unmanned vehicles and personal and service robots enter the mass market, and perform work formerly carried out by humans, we will see many kinds of claims we cannot even imagine today. We will then see courts applying and modifying old liability doctrines for use in cases involving new kinds of unmanned vehicles and robots.
[*] Stephen S Wu is a partner in the Silicon Valley, California law firm Cooke Kobrick & Wu LLP, where he practices business, technology, and intellectual property law. He served as the 2010-2011 Chair of the American Bar Association Section of Science & Technology Law. In addition, he is a lecturer at law teaching technology law at Santa Clara Law School. He can be reached at swu@ckwlaw.com.
[1] 818 N.Y.S.2d 396 (App. Div. 2006), appeal denied, 862 N.E.2d 790 (N.Y. 2007) (‘Jones’).
[2] No. 96-2248, 1997 WL 311586 (8th Cir. Jun. 9, 1997) (per curiam unpublished opinion) (‘Payne’).
[3] 244 Fed. App’x 944 (11th Cir. 2007) (unpublished opinion).
[4] Ibid 948.
[5] Ibid 947.
[6] Ibid 948, 949.
[7] Ibid.
[8] [1984] USCA10 30; 727 F.2d 917 (10th Cir. 1984).
[9] Ibid 921-22.
[10] Ibid 929-30.
[11] Ibid 923-28.
[12] No. 06-2661 (PAM/JSM), 2008 WL 398814 (D. Minn. Feb. 11, 2008) (‘Glorvigen’).
[13] Ibid *5.
[14] Ibid.
[15] Ibid *5-6.
[16] Ibid *6.
[17] See Glorvigen v Cirrus Design Corp, 796 N.W.2d 541 (Minn. Ct. App. 2011).
[18] 943 So. 2d 1204 (La. Ct. App. 2006).
[19] Ibid 1205.
[20] Ibid 1208-09.
[21] Ibid 1209.
[22] 04-CV-5394 (SLT)(RLM), 2008 WL 4724581 (E.D.N.Y. Oct. 24, 2008) (‘Provenzano’).
[23] 92 So. 2d 593 (La. Ct. App. 1957).
[24] Ibid 596.
[25] Redfield v Beverly Health & Rehabilitation Servs, Inc, 42 S.W.3d 703, 710 (Mo. Ct. App. 2001) (affirming denial of defendant’s motion for new trial) (the plaintiff’s decedent died after his ventilator was unplugged and his oxygen tube was disconnected).
[26] 363 Fed. App’x 925 (3d Cir. 2010) (unpublished opinion), cert. denied, 131 S. Ct. 82 (2010).
[27] 83 F. Supp. 373 (N.D. Ohio 1949) (post-trial opinion).
[28] Shaun Fisheries, Inc v Lanegan, Nos. 82-529, 82-687, 82-6148-C, 1983 WL 699 (D. Ore. Sept. 21, 1983) (‘Shaun Fisheries’).
[29] Ibid *2.
[30] Ibid *3-5.
[31] 818 N.Y.S.2d 396, 399.
[32] No. 06-2661 (PAM/JSM), 2008 WL 398814 (D. Minn. Feb. 11, 2008).
[33] 796 N.W.2d 541 (Minn. Ct. App. 2011).
[34] Ibid 549-52.
[35] Ibid 552-58.
[36] Ibid 554 (quoting Alsides v Brown Inst, Ltd, 592 N.W.2d 468, 472 (Minn. Ct. App. 1999)).
[37] 263 Cal. App. 2d 742 (1968).
[38] Ibid 744-45.
[39] Ibid 746.
[40] Ibid.
[41] Ibid 747.
[42] Nos. 82-529, 82-687, 82-6148-C, 1983 WL 699 (D. Ore. Sept. 21, 1983).
[43] No. 23466, 2007 WL 1695109 (Jun. 13, 2007) (unpublished opinion) (‘Miller’).
[44] 692 N.E.2d 195 (Ohio 1998) (per curiam) (‘Scott Fetzer’).
[45] 597 S.E.2d 863 (S.C. Ct. App. 2004) (‘Edens’).
[46] For instance, if, in Scott Fetzer, the employee had sued the manufacturer of the robotic die machine, the court may have found that the employer’s modification of the machine to remove safety guards was a superseding cause of the employee’s injuries.
[47] Mracek v Bryn Mawr Hosp, 363 Fed. App’x 925, 927 (3d Cir. 2010) (unpublished opinion), cert. denied, 131 S. Ct. 82 (2010).
[48] 1997 WL 311586, *2, *3.
[49] 2008 WL 4724581, *5, *6.
[50] No. 92-2227-JWL, 1993 WL 105132 (D. Kan. Mar. 4, 1993).
[51] Ibid *2-*5.
[52] 579 N.Y.S.2d 400 (App. Div. 1992).
[53] See also Royal Ins Co of Am v Crowne Investments, Inc, 903 So.2d 802 (Al. 2004) (reversing denial of motion to set aside default judgment after Royal failed to receive service of process when its robotic mail system malfunctioned; malfunction of the system provided a reason why Royal’s default was excused). If the judgment against Royal had been upheld, and the robot’s malfunction led to a default judgment, it is difficult to see how a court could find that a defect in the robot proximately caused a default judgment. Such a result is not reasonably foreseeable for a general-purpose mail machine.
[54] In addition, the Michigan Supreme Court held that summary judgment was appropriate on a manufacturer’s claim for indemnity against an employer after an accident caused by a malfunctioning robotic machine, Williams v Litton Sys, Inc, 449 N.W.2d 669 (Mich. 1989). Given the exclusivity of the workers’ compensation system, and the lack of an express indemnity in the agreement between the manufacturer and the employer, claims that the manufacturer was entitled to implied indemnity for the employer’s failure to train the worker were barred.
[55] No. 04-1261-HA, 2007 WL 987452 (D. Ore. Mar. 30, 2007) (‘Behurst’).
[56] Ibid *6.
[57] 42 S.W.3d 34 (Tenn. 2001).
[58] 818 N.Y.S.2d 396, 398 (App. Div. 2006), appeal denied, 862 N.E.2d 790 (N.Y. 2007).
[59] [1988] USSC 145; 487 U.S. 500 (1988) (‘Boyle’).
[60] See ibid 512.
[61] See ibid.
[62] In re Aircraft Crash Litig. Frederick, Md., May 6, 1981, 752 F. Supp. 1326 (S.D. Ohio 1990), aff’d sub nom. Darling v Boeing Co, [1991] USCA6 1117; 935 F.2d 269 (6th Cir. 1991) (unpublished opinion affirming for the reasons set forth in the district court’s opinion).
[63] See 752 F. Supp. at 1332-33.
[64] Ibid 1350-58, 1369-71.
[65] Ibid 1370.
[66] 751 F. Supp. 541 (D. Md. 1990).
[67] Ibid 544-45.
[68] No. 04-1261-HA, 2007 WL 987452 (D. Ore. Mar. 30, 2007).
[69] 711 F. Supp. 1377 (N.D. Tex. 1989) (‘Adams’).
[70] Ibid 1379-83.
[71] Ibid 1383.
[72] 651 N.W.2d 383 (Mich. 2002).
[73] No. CV-F-05-203 OWW/SMS, 2007 WL 2898675 (E.D. Cal. Sept. 28, 2007) (‘Bou-Matic’).
[74] No. 3:03CV00515–L, 2003 WL 21517851 (N.D. Tex. Jun. 30, 2003) (‘Rodriguez’).
[75] 604 F. Supp. 630 (W.D. Tex. 1985).
[76] See ibid 634-35.
[77] Ibid 633-37.
[78] Ibid 637.
[79] 506 F. Supp. 1361 (E.D. Wis. 1981).
[80] Ibid 1363.
[81] Ibid.
[82] Ibid 1363-65.
[83] [1984] USCA3 1403; 746 F.2d 977 (3d Cir. 1984).
[84] Ibid 978-79.
[85] Ibid 979.
AustLII:
Copyright Policy
|
Disclaimers
|
Privacy Policy
|
Feedback
URL: http://www.austlii.edu.au/au/journals/JlLawInfoSci/2012/15.html