In this blog post, we will compare the situation to cases involving designated drivers to explore how liability can be interpreted in the event of a traffic accident involving a self-driving car. We have edited and proofread the content to ensure it is natural and easy to read.
The Debate Over Liability for Autonomous Vehicles and Two Perspectives
Scientific progress has always sparked new controversies. During the Industrial Revolution, there were workers who destroyed machines because they believed they were taking away their jobs, and there were religious viewpoints that denied the fact that the Earth was round. Even today, the situation is not much different. While the pace of technological advancement has accelerated, institutions and social perceptions often fail to keep up. Therefore, there is a need for more active social discourse on the latest technologies. Among them, the technology facing the most direct and critical controversy is undoubtedly the self-driving car. Unlike other technologies, self-driving cars possess physical power capable of threatening not only their own operation but also the lives of others.
As the commercialization of self-driving cars draws nearer, the range of controversies surrounding them is also diversifying. Notable examples include ethical decision-making, the potential for terrorist exploitation, hacking, and the issue of liability addressed in this article. Here, “liability” refers to the legal and social responsibilities that arise among victims, insurance companies, drivers, and manufacturers when an accident occurs during autonomous driving. Since artificial intelligence is not perfect, the possibility of accidents cannot be completely ruled out; given these risks, legal frameworks and social consensus must be established before widespread adoption. Accordingly, this article will examine two representative perspectives on liability and develop the discussion by referencing legal interpretations of similar cases. However, the discussion here is premised on fully autonomous driving, where driver intervention is effectively unnecessary.
The first perspective on liability views trust in autonomous vehicles as a form of trust in a warranty. According to this view, even if the driver or others suffer harm in an accident, the decision to purchase the vehicle and use its autonomous driving features was made by the driver, and therefore the driver must bear the consequences. This logic is analogous to not holding a lottery machine responsible for a losing ticket simply because the ticket was purchased via an automated machine, or to the fact that a defense attorney is not held responsible for the outcome of a trial merely because the client retained their services. In other words, the argument is that responsibility for the outcomes of autonomous driving ultimately lies with the user.
The second perspective on liability views trust in autonomous vehicles as a simultaneous delegation of rights and responsibilities. It holds that when a driver delegates the authority to drive to artificial intelligence, the responsibility for safety is also transferred to the autonomous driving system—and, by extension, to the manufacturer that designed it. This is similar to how, when flying, passengers entrust the authority to fly to the airline and the pilot but, in return, are guaranteed responsibility for their safety. This responsibility is generally upheld through a contract of carriage and insurance.
The key difference between these two perspectives ultimately lies in the presence or absence of safety guarantees—that is, the existence of insurance. In the first example as well, if a machine guaranteed a winning ticket or a defense attorney guaranteed the outcome of a trial, the assessment of liability would differ. Similarly, if airfare did not include insurance or a system for compensation for damages, the method of compensation in the event of an accident would inevitably differ. Therefore, the existence of insurance is expected to be a crucial factor when interpreting liability for autonomous vehicles.
Comparison of Liability Structures Using the Designated Driver Case
These issues involve a complex interplay of insurance, liability, and rights, presenting a new challenge to the current legal system, which has not yet fully prepared for the era of autonomous vehicles. For the law to adapt to new technologies, changes across the entire existing legal framework are necessary. Therefore, to examine the issue of liability for autonomous vehicles, it is necessary to examine how similar cases are interpreted under current law. A representative example is an accident involving a designated driver. In the context of designated driver services, liability in the event of an accident is distributed among four parties: the designated driver agency, the designated driver, the vehicle owner, and the victim.
First, the vehicle owner may claim compensation for damages suffered from the designated driver agency and may also demand compensation from the designated driver based on liability for tort. This interpretation is grounded in Supreme Court precedents, the Act on the Guarantee of Compensation for Automobile Damage, and the Civil Code. In other words, even when a vehicle owner hires a designated driver, a certain level of legal protection regarding their own safety is guaranteed.
Meanwhile, the victim may claim damages from the vehicle owner under the principle of employer liability and, at the same time, may also demand damages from the designated driver. In such cases, if the designated driver is insured, compensation may be provided through the insurance; however, if the driver is uninsured or lacks sufficient financial capacity to pay, the vehicle owner may bear significant liability for damages even if they are not at fault. In other words, even if the actual fault lies with the designated driver, the practical liability for compensation may fall on the vehicle owner.
Legal Application and Limitations Regarding Autonomous Vehicle Accidents
To apply this legal interpretation to the issue of liability in autonomous vehicle accidents, one can think of replacing the “designated driver” in existing designated driving cases with the “manufacturer.” In effect, the roles of the designated driver service company and the designated driver are consolidated into the manufacturer. This is because if one simply replaces the “designated driver” with the “autonomous vehicle” itself, there would be no legally identifiable liable party, which could make it difficult to secure appropriate legal compensation. Interpreting the matter in this way yields the following liability structure:
The vehicle owner may claim damages from the manufacturer based on the Civil Code and the Act on the Guarantee of Compensation for Motor Vehicle Damage. Furthermore, the victim may claim damages from the vehicle owner under the provisions on owner’s liability, and from the manufacturer under tort liability and the Act on the Guarantee of Compensation for Automobile Damage. Since the amount of compensation the vehicle owner must bear may vary depending on whether the manufacturer has taken out relevant insurance, the manufacturer needs to secure appropriate insurance coverage in preparation for autonomous vehicle accidents. Alternatively, a system should be established that allows consumers to verify and choose whether insurance has been purchased. However, a key feature is that, regardless of liability for compensation, the manufacturer bears the liability for negligence in the accident.
Of course, this interpretation cannot always be applied as is. This is because it is only valid when it is natural to equate the designated driver and the designated driving service company with the autonomous vehicle and the manufacturer, respectively. In this regard, several limitations exist.
First is the issue of interpreting the term “other person” as defined in the Act on the Guarantee of Compensation for Automobile Damage. Previously, the Supreme Court ruled that if a vehicle owner was riding in the vehicle alongside a chauffeur, the owner would be considered an “other person” because they did not exercise substantive control over the operation of the vehicle. However, it is not yet clear whether a vehicle owner riding in an autonomous vehicle can similarly be regarded as a mere passenger. Nevertheless, even if they are not recognized as a mere passenger, the possibility of claiming damages under tort liability under the Civil Code remains.
Second is the difference in financial capacity to pay damages. While ride-hailing drivers are often uninsured or lack sufficient financial capacity to pay damages, manufacturers are generally corporations with sufficient financial capacity. Therefore, if the manufacturer is held liable, the victim may be able to claim damages from both the manufacturer and the vehicle owner, regardless of whether the manufacturer is insured. In such cases, institutional procedures need to be established to adjust the compensation ratio by considering insurance coverage and the scope of liability.
Third is the issue of the legal status of artificial intelligence and the scope of legal application. Under the current South Korean legal system, artificial intelligence is not recognized as a natural person—that is, a subject of legal rights and obligations. Therefore, this article interprets liability by substituting the manufacturer for the driver based on the Product Liability Act. However, if the legal framework evolves in the future such that AI acquires an independent legal status or is recognized as having legal capacity, the current legal interpretation may also change. Furthermore, whether manufacturers and courts will deem autonomous vehicle accidents to be defects in the product remains a key issue.
These limitations extend beyond the mere application of existing laws to encompass philosophical and ethical debates. Typical examples include questions such as to what extent a self-driving car user is deemed to have actually controlled the vehicle’s operation, what criteria should link liability for negligence to liability for damages, and how the legal status of artificial intelligence should be defined. These issues need to be gradually resolved through social consensus. Since the law is a social norm that reflects the values and ethics of contemporary society, harmony with new technologies must also be achieved through social discourse.
To summarize the above discussion, for the liability for damages in autonomous vehicle accidents to be reasonably allocated, the manufacturer’s insurance coverage is a crucial prerequisite. This is because the scope of what vehicle owners and victims must bear or receive in compensation can vary significantly depending on whether the manufacturer has insurance coverage. Therefore, it is necessary to either mandate appropriate insurance coverage for all autonomous vehicles or establish a system that allows consumers to choose vehicles with insurance coverage. However, given that liability for negligence in accidents rests with the manufacturer, this discussion is, ultimately, intended to serve as an institutional mechanism to clarify liability for compensation.
Of course, for such a system to be established, various issues—including the interpretation of the Act on the Guarantee of Compensation for Automobile Damage, the allocation of liability based on the capacity to pay, and the legal status and scope of application of artificial intelligence—must be thoroughly discussed. Although there are still many challenges to be addressed, this process can be viewed as part of the natural evolution of the law as it adapts to a new technological environment. While the conclusion of this article points to a relatively clear direction—namely, the need for an insurance system for autonomous vehicles—the discussion itself will serve as an important starting point for establishing a new legal framework in a society where artificial intelligence will be increasingly utilized in the future.
