Autonomous vehicles will revolutionize traffic efficiency in the foreseeable future, freeing up hands and reducing traffic accidents. The existing regulations are a solution to the tort liability of traditional motor vehicle traffic accidents, and lack relevant rules on the definition and data collection of autonomous vehicle driving entities, as well as legal provisions on the subject of responsibility, liability and data protection. In 2017, Germany promulgated the Eighth Amendment to the Road Traffic Law to respond to the problem of tort liability after L3 autonomous vehicles were put into the market, accelerate the research and development and implementation of German intelligent vehicle technology, and provide a mirror for the design of tort liability of L3 autonomous vehicles in mainland China from the perspective of comparative law. From the perspective of legislation, the mainland should design the liability for the danger of retaining, refine the liability of the driver, and expand the product liability of the producer, so as to balance the risks and benefits brought about by the development of AI technology.
introduction
With the advancement of autonomous driving technology, the legal challenges it faces have become more prominent, and due to the lack of clear legal provisions, the possible loss of life and property will be immeasurable, leading to the result of various responsible parties passing the buck to each other. In this regard, the legal scholars of various countries have responded at the level of legislative theory. Germany is the first European country to approve commercial autonomous vehicles on the road compared to other countries, creating a favorable environment for the commercial application of autonomous vehicles, has issued a series of regulations and guidelines for autonomous vehicles, and has paid attention to safety and ethical considerations when formulating autonomous driving laws, and explicitly prohibits "ethical dilemmas"; The government is also actively encouraging the innovation and development of autonomous vehicle technology, and allowing new technology companies to enter the open market will help continue to promote technological innovation and legislative maturity. In view of the legal system and the structure of the liability law system, this paper selects Germany as the object of comparative study.
However, many normative documents such as development plans, interim regulations and management regulations issued by the mainland are limited to road tests, and it is difficult to solve all the infringement problems of autonomous vehicles, so there is an urgent practical need to formulate and introduce laws regulating the tort liability of L3 autonomous vehicles in traffic accidents. On this basis, the establishment of a law that can solve the problem of driverless cars in the era of high intelligence is a problem that cars will face after they are fully intelligent. The experience of the German Road Traffic Act (Eighth Amendment) can provide a framework for defining the ethical norms of autonomous vehicles for the design of L3 autonomous vehicles in China, help clarify the connotation of basic issues such as driving subjects, driving rights and obligations, and clarify the characterization of the risk liability of the owner, so as to pave the way for the construction of a tort liability system under the risk society after the arrival of the intelligent era.
1. Examine the risks: Level 3 autonomous vehicles cannot adapt to the traditional tort law system
L3 autonomous vehicles have the situation of alternating control of the human driver and the autonomous driving system, and it is necessary to comprehensively consider multiple factors such as the operation of the human driver and the autonomous driving system, the ownership of the control, and the subject of the accident damage to determine the bearer of the accident liability, which makes it difficult to apply the traditional motor vehicle traffic accident liability and product liability.
(1) The dilemma of the application of traditional motor vehicle traffic accident liability
Level 3 autonomous vehicles are a turning point for driving control agents and driving duty of care. Whether the original principle of attribution can adapt to the legal risks inevitably brought about by the development of artificial intelligence technology is an unresolved problem. The fault judgment standard in the civil law system of the mainland lags behind, and it cannot solve the fault judgment after the autonomous driving system is added to the driving control subject. There is a black box problem in the autonomous driving system, which makes it difficult to apply the principle of no-fault liability.
1. The problem of imputing responsibility under Article 76, Paragraph 1 of the Road Traffic Safety Law
Paragraph 1 of Article 76 of the Road Traffic Safety Law of the People's Republic of China stipulates that fault liability applies to traffic accidents between motor vehicles. Article 91 of the Regulations for the Implementation of the Road Traffic Safety Law stipulates that the traffic management department of the public security organ shall determine the responsibility of the party concerned for traffic accidents according to the seriousness of the party's conduct and fault. In view of this, the determination of traffic accident liability needs to evaluate the force of the party's behavior and the severity of the fault. The evaluation of the magnitude of the force is an objective evaluation of the strength of the causal relationship between the behavior of the parties and the result of the damage. The evaluation of the severity of the fault is an evaluation of the subjective elements of the actor's intention or negligence. The "behavior" of the autonomous driving system includes the timely collection and processing of accident data information in the event of a traffic accident, and the design and construction of this technology is designed and constructed from the perspective of economic benefits to improve traffic efficiency, help motor vehicle human drivers reduce the cost of accident prevention, reduce the duty of care, and balance the value trade-off between the right of way and human rights. However, according to the SAE and NHTSA's criteria for distinguishing different degrees of autonomous vehicle automation, it can be seen that Level 3 is at the watershed between driver assistance and autonomous driving mode. In the road driving stage, there is a situation where the human driver and the driving system cooperate with each other and control the autonomous vehicle alternately, and the vehicle that turns on the autonomous driving no longer has the same subject status and duty of care as the vehicle that has not turned on or does not have the automatic driving device. The reason why the principle of fault liability can be applied between traditional motor vehicles is precisely based on the legal basis that the driving subjects are both human beings and have equal subject status. Therefore, in the age of intelligence, it may not be appropriate for autonomous vehicles to continue to apply fault liability. It is difficult to sustain the fault judgment method constructed around human driving behavior within the existing legal framework. There is controversy as to whether autonomous vehicles can be recognized as driving entities under the regulations and their fault liability can be evaluated based on the "behavior" of the system.
2. The problem of attribution of liability under paragraph 2 of Article 76 of the Road Traffic Safety Law
With regard to the provision of paragraph 2 of Article 76 of the Road Traffic Safety Law of the Mainland, that is, "if the loss of a traffic accident is caused by a non-motor vehicle driver or pedestrian intentionally colliding with a motor vehicle, the motor vehicle party shall not be liable for compensation", the mainstream view in the academic circles is that the motor vehicle party is not at fault liability, and even if the duty of reasonable care is performed, it needs to bear the liability for the victim's losses. Motor vehicles are a source of danger, and it is necessary to pay attention to the protection of the victim's right to life and property, and it is a typical no-fault liability clause for motor vehicles to bear 10% of the compensation liability even if they are not at fault. At the heart of the application of the principle of no-fault liability is the existence of a causal link. That is, the victim needs to prove that the harmful consequences were caused by the infringer's conduct, and there is a causal relationship between the harmful consequences and the act. Autonomous vehicles rely on perception systems and control systems, and it is difficult to determine which component or system the fault is caused by the failure, and whether the error or failure of these systems may lead to traffic accidents. According to the judgment method of the equivalent causality theory, it is difficult to determine whether the infringement is committed by a human driver or an autonomous driving system in the human-computer interactive driving mode. According to this line of thinking, when a human driver intervenes in the system's driving behavior by acting after receiving a warning from the system, but the accident still occurs, it will be impossible to accurately determine which subject is the cause of the traffic accident caused by the autonomous vehicle due to the "black box problem".
(2) The dilemma of the application of traditional product liability
From a normative point of view, product liability is premised on the existence of product defects, and the producer or seller shall bear the damage caused to others due to product defects. It is difficult to determine whether the system manufacturer of L3 autonomous vehicles (hereinafter referred to as autonomous vehicles) can be included in the scope of liability subjects, and the current defect identification standards and exemption reasons for product liability cannot effectively solve the liability regulation of autonomous vehicles as intelligent products.
1. It is difficult to identify the responsible entity of autonomous vehicles
Some of the causes of accidents in autonomous vehicles can be attributed to system vulnerabilities: the system itself has a defect that causes the traffic accident, or the system does not have a defect but the final product is defective due to a defect in the product's integrated design. According to Articles 1202 and 1203 of the Civil Code, if the product is defective and causes damage to others, the manufacturer shall bear tort liability. The victim may seek compensation from the producer or seller. But the manufacturer of the autonomous vehicle and the system manufacturer may not be the same entity. Although this article makes specific provisions on the subject of product liability, it does not specify the scope of producers. In theory and practice, there is a great controversy about the scope of producers, and the first view is that parts manufacturers belong to producers. The second point of view is that since there is a situation where it is impossible to determine which component caused the damage result, and it is not easy to prove it, the subject of responsibility for product liability only includes the producer of the final product. There are differing opinions on whether the manufacturer of parts is a producer under the Product Quality Act, and it is difficult to evaluate whether the actions of manufacturers of autonomous driving systems can be evaluated on the basis of product liability.
2. It is difficult to identify product defects of autonomous vehicles
First, the criteria for determining product liability are too abstract. Autonomous vehicles are different from traditional motor vehicles, and cannot be applied to the national and industry standards of traditional motor vehicles. For the specific judgment of the standard of "unreasonable risk", there are two criteria for determining consumer expectation and risk utility, but there are their own application difficulties. The consumer expectation standard refers to the degree to which the reasonable expectation of the consumer for the product is less than the unreasonable danger of the product, and the consumer has a general social consensus on the product. In general, the consumer is the victim of the product, and the consumer has the general perception of the product when purchasing the product. However, in the case of self-driving car traffic accidents, the victim may not be a consumer, and the victim does not have the same knowledge of the product as the consumer to form an expectation of product design safety. At the same time as the upgrading of artificial intelligence technology, consumers' expectations for the technology and safety performance of autonomous vehicles will increase, and consumers' expectations of standards are subjective, and it is impossible to fully and reasonably expect the artificial intelligence products that are constantly updated with autonomous vehicles, and it is easy to identify too wide on product defects. For the risk utility standard, the concern is whether the product has a more reasonable and safe design, and if it exists but does not adopt it, it is defective. The standard requires the victim to have very strong AI expertise to demonstrate that the algorithm can be programmed with a more secure alternative. In the field of autonomous vehicles, the victim is not the consumer, and it is difficult to obtain the algorithm and data on the producer's side and provide alternative design solutions that outweigh the risks, which can also lead to the risk of data breaches on the part of the producer.
Second, it is difficult to determine the scope of the product liability exemption. Article 41, paragraph 3 of the Continental Product Quality Law stipulates that the producer shall not be liable for compensation if he can prove that "the scientific and technological level of the product at the time of putting the product into circulation is not yet able to discover the existence of the defect". Before autonomous vehicles were put into use, driving systems did not collect traffic data, but as autonomous vehicles are introduced to the market and driving systems are constantly being upgraded, the systems have the autonomy and learning ability to fully collect and analyze external data and information, and the behavioral decisions they make may not be based entirely on the rules inherent in the autonomous vehicles when they are put into circulation, and there are defects that did not exist when they were first put into circulation, and producers can use this to relieve themselves of responsibility. If the standard of "when the product is put into circulation" continues to be applied to the time point of judging the scientific and technological level at this time, it is easy for producers to evade product liability, which is not conducive to motivating producers to continuously improve the error correction mechanism of the autonomous driving system, and is not conducive to the protection of the rights and interests of victims and consumers.
2. Extraterritorial Mirror: A Review of the Revision of the German Road Traffic Act
In 2017, the German Bundestag amended the German road traffic law specifically for Level 3 autonomous vehicles to address the problem of attribution of responsibility for Level 3 autonomous vehicles. The German government encourages the development and innovation of the autonomous driving industry, and in order to solve the ethical problems brought about by artificial intelligence technology, Germany has issued the ethical standard "Code of Ethics for Automated and Connected Motor Vehicle Transportation" (hereinafter referred to as the German "Code of Ethics for Autonomous Driving") to guide autonomous driving legislation before the legislation on autonomous driving, so as to meet the needs of the development and commercial operation of autonomous driving technology.
(1) The amendment to the Eighth Amendment to the German Road Traffic Act
1. Clarify the concept of self-driving cars
The Eighth Amendment to the German Road Traffic Act adds clauses 1a and 1b. The legal status of an autonomous vehicle as a "highly autonomous or fully autonomous vehicle" is affirmed in Article 1a, and the specific meaning of the term "highly autonomous or fully autonomous vehicle" is further explained in paragraph 2: " It can flexibly steer the vehicle and complete longitudinal or horizontally guided driving routes, obey traffic laws, can take over or deactivate from the driver at any time, determine whether it is necessary for the driver to take over the vehicle, notify the driver through multiple senses or other means to take over the vehicle when the autonomous driving program is manipulating the vehicle, and give early warning in case of emergency. ”
2. Clarify the rights and obligations of human drivers
While giving the human driver of an autonomous vehicle the right to freely decide whether or not to turn on the autonomous driving mode, the law stipulates that the human driver may temporarily not take over the vehicle, but at this time he needs to be vigilant and take over. The human driver should be alert and take over in two situations: when the autonomous driving system asks the human driver to take over and when the human driver realizes that the autonomous driving function is not driving as intended.
3. Clarify the data processing rules for autonomous vehicles
Any car equipped with a "black box" autonomous driving system should be able to accurately identify possible hazards, so as to determine which responsible entity bears more legal responsibility. First, clarify how long the data will be kept. The law stipulates that general data will be kept for a period of six months, and for self-driving cars involved in traffic accidents for at least three years. Second, the law stipulates that the position data and time data of the autonomous vehicle determined by the satellite navigation system, the switching data between the human driver and the autonomous vehicle, the data of the human driver taking over the motor vehicle, and the fault data of the autonomous vehicle system should be stored. Finally, the data transmission requirements should be clarified, on the one hand, the autonomous vehicle authority can require the owner of the autonomous vehicle to transmit data, and on the other hand, the owner of the autonomous vehicle can transmit data to a third party in the form of data anonymization.
(2) German L3 autonomous vehicle tort liability system
Germany's tort liability regulation for L3 autonomous vehicles basically follows the traditional traffic accident liability and product liability, and does not completely change the traditional motor vehicle traffic accident tort liability system, and constructs the driver's presumption of fault liability, the owner's danger liability, the manufacturer's product liability, and the producer's fault liability.
1. Presumptive liability for the fault of the driver of the autonomous car
To ensure safety, all self-driving cars must have a human driver and obey traffic laws. The presumption of fault applies to the driver, and in the event of a traffic accident, the driver must prove that he or she has exercised reasonable care in order not to bear tort liability. According to § 823 of the German Civil Code, the driver's presumption of liability for fault is limited to a limit, beyond which the driver is liable for fault.
2. Liability for the dangers of self-driving car owners
Regardless of whether the owner of the self-driving car is liable or not, the victim should be held responsible for the damage caused by the owner's ability to operate the vehicle and the risks that may arise. In addition, the new law adds Article 12, paragraph 1, which increases the amount of compensation for the no-fault liability of car owners, so as to better protect the rights of victims.
3. Product liability for self-driving car manufacturers
In the case of autonomous vehicle manufacturers, there is no direct liability for the failure of the system, which is only a secondary product liability for system failures, and cannot replace or precede the driver's ethical decisions. According to Section 1 of the German Product Safety Act, the manufacturer's liability for the product is no-fault. When a self-driving car is involved in a traffic accident due to a defective product, the victim can claim product liability from the manufacturer under the Act if he or she can prove the causal relationship between the defect and the result of the damage. According to Article 1a, paragraph 2, item 2 of the Eighth Amendment to the German Road Traffic Act, the definition of "highly autonomous or fully autonomous vehicle" includes the characteristics of being able to comply with traffic laws, which means that the software technology of an autonomous vehicle complies with the same traffic laws as the driver of a conventional motor vehicle, and at least the reasonable care required of the driver of a conventional motor vehicle. Therefore, if the autonomous driving system fails to fulfill the reasonable duty of care expected of human drivers and causes a traffic accident, the system can be found to be defective.
4. Special Responsibilities of Autonomous Vehicle Manufacturers
The responsibilities of product manufacturers and producers in Germany are not the same. The manufacturer of the product bears the responsibility for the fault. According to §§ 823 and § 276 of the German Civil Code, the producer has an obligation to ensure that the product fails to meet reasonable expectations due to the damage caused by a defect in the product due to the producer's subjective intention or negligence, the producer has violated the obligation of safety and shall compensate the victim.
(3) The characteristics and significance of the revision of German L3 autonomous vehicles
According to the discussion of German scholars, the amendment does not address many of the key problems of autonomous vehicles, such as the lack of clarity on the distinction between highly and fully autonomous vehicles, the failure to mention the problem of hacking, and the lack of clarity on the warning obligations of human drivers. Most importantly, there is a lack of a system that treats car manufacturers and motor vehicle owners and drivers fairly. For example, in terms of the liability limit, if the accident is caused by a malfunction of the driving system, the liability limit of the car owner is 10 million euros, which is twice the amount compared to the previous law and increases the liability of the car owner, but the driver of the autonomous vehicle does not apply to this limitation of liability.
In general, the amendment is mainly to regulate the legal issues of Level 3 autonomous vehicles, which are in the transition stage of man and machine, and the human driver still has a lot of control over the vehicle, and should support the attribution of more obligations on the owner of the vehicle. While following the traditional theory that one entity bears the responsibility for danger and the other subjects bear the fault liability, the amendment clarifies the legal status of autonomous vehicles from the legal level, clarifies the relationship between the rights and responsibilities of human drivers and the entities of the autonomous vehicle industry, and constructs flexible legal rules within the existing legal framework in Germany, responding to the difficult problems faced by many autonomous vehicles in the field of infringement, and is a normative guide for dealing with tort liability accidents of autonomous vehicles in subsequent judicial practice.
3. Local construction: Improve the legal regulation of tort liability for L3 autonomous vehicles
In the current traffic situation where roads are mixed, it is necessary to build intelligent designs that can turn off the autonomous driving system at any time and have a human driver take over. This design respects the free will of human beings while setting the overall tone for tort liability for autonomous vehicles. Based on the reflection of Germany's L3 autonomous vehicle legislation on China, the mainland should jointly attribute the responsibility of autonomous vehicles to accident liability and product liability through the legislative theory channel under the premise of establishing ethical norms, clarify the liability methods and principles of each responsible entity, and balance the risks and benefits brought by the intelligent era and protect the rights and interests of the weak in combination with the defect identification standards of product liability.
(1) Complete the legal and ethical basis for autonomous vehicles
At present, although there are principled provisions of "protecting personal safety, protecting the property safety and other legitimate rights and interests of citizens, legal persons and other organizations" in some local autonomous driving regulations, there is a general lack of provisions and descriptions of legal and ethical foundations. In contrast, Germany is the first country in the world to issue ethical guidelines for autonomous driving, and the Eighth Amendment to the German Road Traffic Act clearly stipulates the protection of human rights, such as "taking into account the importance of legal interests and giving the highest priority to the protection of human life when causing unavoidable vicarious damage to different legal interests", which is in line with Article 7 of the German Code of Ethics for Autonomous Driving, which "puts human life first". In the process of autonomous driving legislation, the mainland should also strengthen the protection of human rights and personal privacy rights, reflect the respect for human free will, and establish an ethical autonomous driving regulation and control system. In response to the most basic ethical issues that can be involved in autonomous vehicles, the following basic ethical provisions are designed:
1. Safeguard human rights
Protecting human rights, including protecting the life, health and safety of all traffic participants, safeguarding the dignity of traffic participants, safeguarding the free will of traffic participants, upholding equal relations between traffic participants, and giving traffic participants and the public the right to know. The above provisions will also need to be integrated with the autonomous vehicle laws that will be revised or newly established in the future. For example, in future legislation on autonomous driving, the right of human drivers to choose between system driving and human driving modes can be retained. On the one hand, it is in line with the ethical clause of guaranteeing the free will of the participants. On the other hand, in order to implement the principle of equality of all traffic participants, in the context that the ethical norms give the holder more choices and the victim enjoys fewer choices, the tort liability bearing method in which the holder bears more duty of care can be constructed to balance the rights and interests of all subjects and achieve the value orientation of equality for all as much as possible.
2. Risk balancing
Seek to balance the risks between the use of autonomous driving technology and the protection of people. Autonomous vehicles are designed to improve driving safety and revolutionize transportation efficiency, but the laws of modern society cannot completely eliminate risks, only control them appropriately. Article 3 of the German Code of Ethics for Autonomous Driving states: "Accident prevention is the guiding principle, but the remaining risks that are technically unavoidable do not prevent the implementation of autonomous driving, taking into account the basic risk-benefit balance." "In the face of the possible risk scenarios brought about by L3 autonomous vehicles caused by the use of artificial intelligence technology, the responsibilities of all parties can be balanced through legal means, and legislators must pay attention to the rule of law in the control function of risk control to achieve the protection of people.
(2) Respond to basic legal issues related to autonomous vehicles
Basic legal provisions such as the concept and legal nature of autonomous vehicles are the basis for subsequent relevant legislation, and detailed provisions on the data processing of autonomous vehicles are helpful to solve the problem of the allocation of the burden of proof when designing a tort liability system, and also to the determination of causation.
1. Clarify the concept of self-driving cars
Article 1a of the Eighth Amendment to the Road Traffic Act defines highly autonomous vehicles as "capable of complying with the traffic laws that guide the vehicle" and "being able to judge whether it is necessary for the driver to take over the vehicle", reflecting the high duty of care and responsibility of automobile manufacturers, thus leaving room for interpretation of the criteria for determining product liability and the reasons for exemption from liability for automobile manufacturers. Continental can learn from Germany to define autonomous vehicles as "highly autonomous or fully autonomous vehicles", and explain their specific connotation characteristics, and to clarify the difference between highly and fully autonomous vehicles, the specific provisions can be supplemented in Article 119 of the Road Traffic Safety Law, and the legal positioning of autonomous vehicles should be clarified as the legal object while explaining the concept of motor vehicles.
2. Clearly state the rights and obligations of the driving subject
German law provides detailed provisions on the rights and obligations of human drivers, and the duty of care that human drivers should perform includes illegal operation, improper takeover, failure to take over in a timely manner, failure to recognize the need for takeover, etc., and the clarification of such duty of care can help determine the fault of human drivers. The driver's rights include starting and shutting down the autopilot program. In order to clarify the fault liability of mainland drivers, the driver's duty of care can be added to Article 21 of the Road Traffic Safety Law, including but not limited to "the obligation of human drivers to supervise autonomous vehicles, the obligation to accept system warnings, the obligation to inspect the system settings, and the obligation to actively or passively take over the motor vehicle immediately".
3. Clarify the backup and supervision of key data
The data storage time of L3 autonomous vehicles and third-party regulatory agencies should be determined in the form of law, leaving room for flexibility in the continued construction of the law on tort liability for autonomous vehicles. A special chapter "Cyber Security and Data Protection" has been set up in the Road Traffic Safety Law of the People's Republic of China to regulate the supervision and management of cyber security risks of L3 autonomous vehicle products and the content, duration and subject of data protection. If it is specified in detail that the manufacturer of the autonomous driving system needs to provide the autonomous driving data recording device and the installation certificate, the supervision device should also be connected to the third-party supervision platform, and the network security assessment and management mechanism established by the third-party institution shall be used as the standard, and the driving data shall be reported to the relevant state departments in a timely manner when there are national security and personal privacy security issues, so as to ensure the authenticity of the data. Penalties can be imposed if self-driving car manufacturers refuse to open up data and algorithms.
(3) Supplement the legal provisions on tort liability for L3 autonomous vehicles
Although the German amendment does not change the traditional principle of attribution of traffic accident liability and product liability, it has added some requirements for some responsible entities: first, it increases the compensation limit of the no-fault liability of the retainer to protect the rights and interests of the victim and balance the new risks brought about by the intelligent system in the process of joining the driving subject. Second, self-driving car manufacturers need to have at least the same level of care as traditional human drivers. In order to solve the problem of mismatch between the L3 autonomous vehicles and the existing tort liability in the mainland, we can learn from the legislative experience of Germany and combine the national conditions of the mainland to increase the proportion of liability borne by the owners and producers, and construct tort regulations applicable to both traditional motor vehicles and L3 autonomous vehicles in the intelligent era.
1. Establish a liability for the danger of possession
First of all, the internal system of traditional tort law in mainland China is based on the tort of natural person's fault, which cannot be fully applied to the risk society in the era of artificial intelligence. Although the development of self-driving cars can greatly reduce the traffic accident rate, it is still highly dangerous due to the frequent traffic accidents of self-driving cars in recent years, various design flaws and "algorithm black box" problems existing in the system itself. Based on the attributes of autonomous vehicles that meet the risk liability, the three theoretical bases of risk liability in tort law (including risk opening theory, risk control and dispersion theory, Compensation theory) can be seen that the behavior of the owner of the L3 autonomous driving car to buy the vehicle will open the source of danger, and the consumer has a better understanding of the artificial intelligence product of the autonomous vehicle than the average person, and the human driver can take over the vehicle at any time during the driving process, has the ability to control, and also enjoys the various benefits of the consumer, and the owner should pay for the risk.
Second, the scope of the above risks needs to be limited. As no-fault hazard liability, the risks that the owner is responsible for are limited to the risks associated with the operation of the autonomous vehicle, and cannot be broadened to cover all ordinary life risks, such as the driver's own physical reasons, system design defects, third party intentions, and the fault of the hired driver, etc., should be excluded. To sum up, it is most appropriate for the person who initiates the dangerous activity and has a dominant management position and profits from it to be the risk bearer, and this entity is the legal basis that best meets the above-mentioned risk liability when the person is the keeper.
First, Article 76 of the Road Traffic Safety Law stipulates that the party to the motor vehicle shall bear no more than 10% of the liability for compensation, which is in line with the legal theory of no-fault liability, indicating that the principle of attribution of liability is inherently risky. Second, the relevant judicial interpretations of the Supreme People's Court also have precedents for applying the criteria for judging operational interests and operational dominance in hazard liability. Since Level 3 autonomous vehicles and ordinary motor vehicles belong to the mixed driving stage on the road, it is necessary to minimize the impact on the current legal system and localize the no-fault retention system to make it easier to integrate into the existing motor vehicle traffic accident liability. Article 76 of the Road Traffic Safety Law may be amended as follows: "If a motor vehicle is involved in a traffic accident and causes personal injury or death or property damage, the insurance company shall compensate within the limits of the compulsory liability insurance of the third party liability of the motor vehicle; ”
2. Refine the presumption of fault liability of the driver
First of all, it is clear that the driver of Level 3 autonomous vehicles is a human driver. Although the autonomous driving system is also the main body of driving control, it does not have the ability to directly control the driving of the car, and must be connected and interacted with the car parts to achieve the purpose of controlling the driving of the car. Therefore, autonomous driving systems can only be seen as driving assistants or assistance systems. Moreover, the autonomous driving system does not have legal personality, does not have subjective intention and negligence, and cannot be applied to the rule system based on fault liability. It would be more appropriate to define autonomous vehicle systems as products with liability for their applicable product liability.
Second, although the duty of care of the driver of an autonomous vehicle includes the duty of care of the driver of a traditional motor vehicle and the new duty of care, the duty of care of the driver of an autonomous vehicle is actually reduced compared with that of a traditional motor vehicle driver, because the duty of care of the human driver is different in different human-machine driving modes. In the first stage, when the human driver receives a prompt from the system to take over the vehicle, the system controls the vehicle more, and the human driver only needs to bear individual duty of care (such as the obligation to remain vigilant and maintain attention). Adopting the objective standard of duty of care of a reasonable person, the human driver's omission to perform some of the above obligations will constitute a breach of the duty of action, which is illegal and meets the composition of fault, unless the human driver can provide evidence to prove that he has fulfilled his reasonable duty of care and shall bear the presumption of fault liability. In the second stage, when a human driver takes over the vehicle, the human driver does not directly bear all the responsibility for the fault, because the human driver takes over the vehicle with reaction time. In the event of a traffic accident before the transfer of driving authority to a human driver, the above-mentioned method of imputing responsibility before the human driver takes over may apply. If the driver is subjectively intentional or negligent in committing an act of delaying the takeover, he or she shall be presumed to be at fault. In the third stage, when a human driver takes over the driving of the vehicle, the human driver, as the controller of the L3 autonomous vehicle, bears the no-fault liability when the driving subject and the owner are the same subject, and the driving subject and the owner are not the same subject, and the driving subject bears the fault liability.
Finally, the second half of Article 1209 of the Civil Code needs to be changed to "the owner and manager of the motor vehicle shall be liable for compensation for the occurrence of damages" in the corresponding amendments, so as to make it clear that when the owner of the car is inconsistent with the driver, the driver shall bear the fault liability and the owner shall bear the no-fault liability.
3. Expand the product liability of the producer
First, it is clear that the producer of Level 3 autonomous vehicles is the subject of product liability. There may be multiple manufacturers of products within a self-driving car, but there is a final producer who assembles the product, labels it, and puts the product on the market, and this final producer is responsible for the finished form of the product, as well as for each defect in the final product. If an autonomous vehicle meets the constituent elements of a product defect, the final producer must bear product liability. As a manufacturer of some products in Level 3 autonomous vehicles, the defects of autonomous driving system manufacturers are of course the key defects in the entire autonomous vehicle, which will have a crucial impact on the operation of other components and the operation of the entire motor vehicle.
Second, the first is to combine the "risk utility standard" and the "consumer expectation standard" to determine the product defects of L3 autonomous vehicles. On the one hand, this criterion conforms to the legal principle of consistency between subjectivity and objectivity. If the consumer expects a high standard, the product risk is greater in the eyes of the consumer and the cost of improving the existing product is correspondingly higher, and if the consumer expects a low standard, the product risk is less in the eyes of the consumer and the cost of improving the existing product is also lower. There is a positive correlation between consumer expectations that affect the judgment of existing risks and the cost required for future improvements. Therefore, when judging whether a product is defective, it should be judged in combination with the subjective consumer expectation standard and the objective risk utility standard. On the other hand, the standard is also in line with the ethical norms that guarantee human rights and balance risks. The data leakage that may result from the application of the risk utility standard in product defect judgment is one of the main manifestations of AI data ethical risks. In order to balance the relationship between risk and return, it is necessary to combine the consumer expectation standard and the risk utility standard. Because the higher the reasonable expectations of consumers, the higher the safety standards of manufacturers for manufacturing L3 autonomous vehicles, the improvement of safety is in line with the ethical basis for protecting human rights, and at the same time can promote the continuous improvement of autonomous driving technology, reduce the risk of data leakage, and comply with the ethical code of risk balance. Second, it is necessary to adjust the time point for judging the level of science and technology. Based on the technical characteristics of AI products with autonomy and learning ability, the duty of care on the part of the producer will be extended from before the product is provided to after the product is provided. In the reasons for the exemption from product liability on the producer's side, the standard of "when the product is put into circulation" should not continue to be applied to the time point of judging the scientific and technological level, but can be changed to the standard of "when the system is upgraded after the product is put into circulation", so as to make the product of autonomous vehicles safer.
Finally, when the law is amended, it is necessary to establish in the Product Quality Law that the criteria for determining product defects of Level 3 autonomous vehicles are to combine the "risk utility standard" and the "consumer expectation standard". By dividing the three time periods of "before the user of the motor vehicle receives the prompt from the system to take over the vehicle", "during the process of the user taking over the vehicle" and "after the user of the motor vehicle takes over the vehicle", it is clarified that when the autonomous driving system enjoys the driving authority and the human driver is not at fault but fails to take over in time, the manufacturer is liable for the infringement of the defective product.
conclusion
On the one hand, there is a certain degree of mismatch between the specific provisions and liability assumption methods of Article 76 of the Road Traffic Safety Law and the determination of liability in the event of a traffic accident by autonomous vehicles, and on the other hand, the current institutional arrangement of product liability in terms of the subject of responsibility and product defects cannot effectively solve the problem of liability regulation of autonomous vehicles, which is an intelligent product. The German Code of Ethics for Autonomous Driving lays the ethical foundation for the development of German autonomous driving law and is the key to the rapid commercialization of German autonomous vehicles. The Eighth Amendment to the German Road Traffic Act responds to the difficult question of how to attribute L3 autonomous vehicles: on the one hand, Germany has not added new traffic accident liability and product liability rules, because its existing laws can solve the problem of motor vehicle infringement; Liability attribution and other issues are stipulated, and the maximum amount of compensation for personal injury and property damage caused by self-driving vehicles is increased, which solves some legal problems in the application of artificial intelligence in the field of motor vehicles.
The experience of the German autonomous driving law is helpful to solve the construction problem of the continental autonomous driving law. First of all, it is necessary to construct an ethical code of AI by answering the most basic ethical questions of autonomous vehicles, which should be used as the philosophical basis for the legislation of tort liability of autonomous vehicles. Second, with the popularization of autonomous driving technology and changes in regulatory policies, ethical issues will also have a significant impact on tort liability. Whether the human driver should bear more responsibility, or the automobile manufacturer should bear more responsibility based on the control of the autonomous vehicle by the system, and what kind of responsibility, the basic legislative matters of L3 autonomous vehicles should be responded to in the existing civil code, product quality law, and road traffic safety law according to the national conditions of the mainland, so as to cooperate with the revised design of L3 autonomous vehicles in which the owner and producer bear more tort liability. Finally, considering the development potential of autonomous vehicles, the philosophical determination of the ethical issues of AI robots may be the solution for the subsequent practical and disciplinary development of intelligent products.